Steps to Microsoft 365 Copilot Extensibility with Gautam Sheth [MVP]
![Steps to Microsoft 365 Copilot Extensibility with Gautam Sheth [MVP] Steps to Microsoft 365 Copilot Extensibility with Gautam Sheth [MVP]](https://images-cf.getpodpage.com/cdn-cgi/image/quality=70,fit=contain,format=auto,width=480/https://images.podpage.com/tr:w-1200,h-630,cm-pad_resize,bg-blurred_70/https://img.youtube.com/vi/R4flzbE6pEA/maxresdefault.jpg)
![Steps to Microsoft 365 Copilot Extensibility with Gautam Sheth [MVP] Steps to Microsoft 365 Copilot Extensibility with Gautam Sheth [MVP]](https://images-cf.getpodpage.com/cdn-cgi/image/quality=70,fit=contain,height=130,format=auto,dpr=1/https://img.youtube.com/vi/R4flzbE6pEA/maxresdefault.jpg)
In this episode of M365.fm, host David Warner speaks with MVP Gautam Sheth about Microsoft 365 Copilot extensibility and how organizations can go beyond the out-of-the-box Copilot experience. Gautam explains the different extensibility options available, including declarative agents, custom engine agents, Copilot connectors, and plugins, helping listeners understand when and why each approach should be used.
The discussion focuses on practical steps for extending Copilot with organizational knowledge, business processes, and external systems to deliver more relevant and actionable AI experiences. Gautam highlights the importance of grounding Copilot in enterprise data, leveraging Microsoft Graph, and designing solutions that align with real business scenarios rather than simply adding AI for its own sake.
The episode also covers governance, security, and adoption considerations, emphasizing that successful Copilot implementations require a balance between technical capabilities and user needs. Listeners gain insights into the current state of the Microsoft 365 extensibility ecosystem, common implementation patterns, and recommendations for developers and architects looking to build intelligent, enterprise-ready Copilot experiences.
Overall, the conversation provides a practical roadmap for understanding and implementing Microsoft 365 Copilot extensibility, making it valuable for developers, architects, and IT leaders exploring how to maximize the value of AI within Microsoft 365.
Have you ever wished you could automate repetitive tasks or create a workflow that fits exactly how you work? With microsoft 365 copilot, you gain tools that let you do just that. You can use copilot to draft emails in Outlook, generate reports in Excel, or assign tasks in Teams—saving hours each week. The shift to a cloud-first architecture and deep integration with Microsoft Graph means you experience seamless data flow across all your microsoft 365 apps. As a result, you see less manual work, faster workflows, and more time for what matters most.
| Team | Metric | Real-world Benchmark |
|---|---|---|
| Finance | Reporting cycle time | Reduced from 2 days to 4-6 hours (75% reduction) |
| Legal | Contract review time | 4 hours saved per week per legal professional |
| Marketing | Brief creation time | Reduced from 3 hours to 30-60 minutes |
| General | Productivity value | 3 hours saved per week per employee |
You can build on the existing microsoft 365 copilot stack or personalize your own solutions. Imagine how custom tools could reshape your daily tasks and help you reach new levels of productivity with microsoft and copilot.
Key Takeaways
- Microsoft 365 Copilot automates repetitive tasks, saving you hours each week.
- You can customize Copilot to fit your unique workflow, enhancing productivity.
- Integrate third-party apps like Salesforce and Jira to streamline your processes.
- Use low-code tools to create custom agents without needing advanced technical skills.
- Automate document creation and email drafting to focus on more important tasks.
- Leverage data analysis features to make informed decisions quickly.
- Regularly assess security and maintain your custom tools for optimal performance.
- Engage with community resources for support and to share your own solutions.
Microsoft 365 Copilot Extensibility
What Is Extensibility
You can think of copilot extensibility as the ability to customize and expand what microsoft 365 copilot can do. This means you are not limited to the default features. You can add new tools, connect to outside data, and create solutions that fit your unique needs. Extending microsoft 365 copilot lets you build on top of the microsoft 365 platform, so you can automate tasks, personalize workflows, and connect with other apps or services.
Here is a quick look at some of the main extensibility features available in microsoft 365 copilot:
| Feature | Description |
|---|---|
| editorial_answers | Lets agents match user queries to predefined Q&A pairs. |
| default_response_mode | Sets how agents respond: Auto, Think deeper, or Quick response. |
| depends_on | Specifies dependencies for conversation starters. |
| New agent templates | Eight templates for common workplace scenarios, like AI Learning Advisor. |
| Evaluation framework | Tools to refine agent performance, such as the Agent Evaluations CLI tool. |
| Package Management API | Allows IT admins to manage apps and agents, including blocking packages. |
| Agent Registration API | Enables programmatic registration and management of agents. |
| Copilot policy settings API | Unified endpoint to read and update Copilot settings across policies. |
Core Purpose
The core purpose of copilot extensibility is to help you get more value from microsoft 365 copilot. You can use extensibility to automate your business processes, connect to external data, and create custom agents that solve real problems. By extending microsoft 365 copilot, you can make your daily work easier and more efficient. You can use tools like Copilot Studio or Visual Studio Code to build plugins that fit your workflow. These plugins can automate repetitive tasks, connect to APIs, and bring your business data into copilot.
Tip: When you use copilot extensibility, you can create agents that reflect your organization’s strengths and needs. This helps you stand out and work smarter.
Empowering Users
Copilot extensibility puts you in control. You can create custom solutions that match your team’s workflow. With microsoft 365 copilot, you can use declarative agents, custom engine agents, and connectors to expand what copilot can do. You can personalize how copilot interacts with you, add new instructions, and even explore information outside the microsoft 365 data estate. This means you can unlock new workflows and take action from one simple interface.
You do not need to be a developer to benefit from extending microsoft 365 copilot. Many tools offer clear entry points and templates, so you can start building right away. If you want to go further, you can use advanced tools like Copilot Studio Agent Builder or Azure AI. These options let you create agents with different levels of complexity, so you can solve both simple and complex challenges.
By using copilot extensibility, you can make microsoft 365 work the way you want. You can automate, personalize, and connect—giving you more time to focus on what matters most.
Custom Tools for Productivity

You can unlock new levels of productivity with microsoft 365 copilot by using custom tools that automate tasks, integrate apps, and personalize workflows. Copilot extensibility gives you the power to build solutions that fit your needs. You can use the microsoft 365 agents toolkit, connectors, and APIs to create agents that streamline your daily work. These tools help you save time, reduce manual effort, and focus on important tasks.
Automating Tasks
You can automate many tasks with copilot. Custom agents make it easy to handle repetitive work and boost efficiency. Here are some common use cases:
- Talent sourcing, screening, and acquisition: Automate candidate screening and email correspondence.
- New hire onboarding: Personalize onboarding experiences and automate training processes.
- Analyzing employee performance: Identify training needs by analyzing performance data.
- Email composition and management: Suggest phrases, summarize points, and organize emails.
- Document creation and editing: Draft and format documents quickly.
- Spreadsheet formulas and analysis: Simplify spreadsheet tasks and suggest formulas.
- Calendar scheduling and meeting management: Streamline scheduling and meeting coordination.
- Project planning and task management: Offer insights and reminders for project management.
- Data analysis and visualization: Extract insights from data.
- Presentation design and delivery: Enhance presentation design and provide feedback.
- Document translation and localization: Facilitate translation and localization of documents.
- Content creation and curation: Accelerate content creation and curation processes.
Document Creation
You can use copilot to draft, summarize, and rewrite documents in Word. Custom agents help you create detailed documents using plain text prompts. You can automate formatting and editing, making document creation faster and more consistent. In Excel, copilot assists you with data analysis, formula generation, and chart creation. You can ask copilot to prepare a chart based on raw data, saving time and reducing errors.
Tip: Automating document creation with copilot extensibility lets you focus on content instead of formatting.
Calendar Management
Copilot helps you manage your calendar and meetings. You can automate scheduling, meeting coordination, and reminders. Custom agents can summarize meeting discussions and outline action items for follow-up in Teams. You can streamline your calendar by letting copilot handle invitations and conflicts. This reduces manual work and keeps your schedule organized.
Note: Many organizations report significant time savings by automating calendar management with microsoft 365 copilot.
Integrating Apps
Copilot extensibility allows you to connect microsoft 365 with third-party applications. You can use connectors and APIs to synchronize data and automate workflows across platforms. The microsoft 365 agents toolkit makes integration simple, even for users with low-code skills.
Third-Party Connections
You can integrate copilot with Salesforce, SAP, and Jira. Here is how you can connect these apps:
- Enable Microsoft Dataverse Connector to synchronize data between Salesforce and Power Platform.
- Use Power Automate connectors for Salesforce to automate processes like summarizing contracts and generating reports.
- Integrate copilot in Teams or Dynamics 365 by embedding Salesforce data through API connections.
- Train copilot to recognize Salesforce-specific queries and generate responses based on CRM data.
- Enable Power Platform SAP connector to connect SAP with the platform.
- Trigger workflows with Power Automate.
- Use copilot in Power BI to fetch and analyze SAP ERP data.
- Deploy Azure Logic Apps for advanced API-based integration with SAP.
- Install the Jira Cloud Connector in Power Automate to synchronize Jira tasks with microsoft 365.
- Embed copilot in Outlook or Teams to retrieve Jira task updates on request.
- Use copilot in Power BI to generate Jira project reports and insights into sprints.
- Customize workflows in Jira-specific queries and automate their responses.
Team Collaboration
You can use copilot to enhance collaboration in Teams, SharePoint, and Outlook. Custom agents summarize meetings, list action items, and automate document movement based on user-defined rules. You can organize notes in OneNote and create summaries for better information management. Copilot helps you build custom reports in Power BI and track KPIs in real time. You can automate CRM and ERP actions in Dynamics 365, such as summarizing customer interactions and suggesting next steps for sales managers.
| Application | Use Cases | Example |
|---|---|---|
| Word | Draft, summarize, rewrite documents | Draft detailed documents using plain text prompts. |
| Excel | Analyze data, create formulas, generate charts | Ask copilot to prepare a chart based on raw data in an Excel spreadsheet. |
| PowerPoint | Create presentations | Generate presentations with data analysis and visualizations. |
| Outlook | Summarize threads, draft emails | Summarize email threads and draft responses efficiently. |
| Teams | Summarize meetings, list action items | Summarize meeting discussions and outline action items for follow-up. |
| OneNote | Organize and summarize notes | Organize notes and create summaries for better information management. |
| SharePoint | Document management and collaboration | Automate document movement based on user-defined rules. |
| Power BI | Build custom reports | Create interactive charts to track KPIs in real time. |
| Dynamics 365 | CRM and ERP actions | Summarize customer interactions and suggest next steps for sales managers. |
Personalizing Workflows
Copilot extensibility lets you personalize workflows to match your business rules. You can use low-code tools to customize microsoft copilot and create agents that align with your needs. You can automate tasks and improve responsiveness by integrating AI into your existing workflows.
Custom Templates
You can create custom templates for documents and emails in microsoft 365 copilot. These templates help you maintain consistency and efficiency. You can build branded email templates, standardized financial report templates, and uniform PowerPoint presentations. Custom templates make document creation faster and ensure your team follows the same standards.
Tip: Custom templates in copilot save time and help your team stay organized.
Data Analysis
You can use data analysis tools integrated with microsoft 365 copilot to gain insights and make informed decisions. Copilot provides trend analysis, predictive analytics, and custom performance dashboards. You can adjust workflows based on analytical insights and improve your business outcomes. Copilot extensibility supports personalized workflow adjustments and helps you respond quickly to changing needs.
Note: Many organizations report increased productivity and faster decision-making by using copilot for data analysis.
You can see real-world results from organizations that use microsoft 365 copilot custom tools. The City of Burlington reduced permit approval time from 15 weeks to 5-7 weeks. Finastra cut campaign time-to-market from three months to less than one. GoTo Group saved over seven hours per week per engineer. Milpark Education improved student support efficiency by 50%. Teladoc Health saved five hours per week per employee. Uniper increased auditor productivity by 80%. Vodafone saved employees an average of three hours per week in early trials. Globo achieved a 90% reduction in time required to validate financial data and a 97% reduction in execution time for complex spreadsheet tasks.
You can use copilot extensibility, custom agents, and the microsoft 365 agents toolkit to automate, integrate, and personalize your workflows. You can create agents with low-code tools or dive into agent development for advanced solutions. Building agents with copilot gives you the flexibility to solve unique challenges and drive enterprise ai innovation in your organization. Microsoft 365 copilot and extensibility help you work smarter, save time, and achieve more.
Microsoft 365 Copilot Integration Options

Microsoft 365 Copilot gives you several ways to connect your tools, data, and workflows. You can use copilot extensibility to build solutions that fit your business. These integration options help you automate tasks, access information, and create agents that work across the Microsoft 365 environment.
Agents in Copilot Studio
Agents in Copilot Studio act as specialized AI assistants. You can create agents that focus on specific business needs. These agents use organizational knowledge and automation to make your work easier. For example, you can build an agent that retrieves information, summarizes data, or sends emails. You have two main ways to build agents: a declarative approach or a custom engine approach. The core parts of each agent include knowledge and actions. These parts help you automate business processes. You can add more layers, such as orchestrators or user experience features, to make your agents even smarter. Copilot extensibility lets you create agents that match your workflow, whether you use low-code tools or dive into agent development.
Tip: Use the microsoft 365 agents toolkit to start building agents quickly and easily.
Connectors and APIs
Connectors and APIs help you bring data from outside sources into Microsoft 365 Copilot. You can use these tools to connect with systems like ERP, CRM, or other business apps. Microsoft Graph connectors serve as a bridge between Microsoft 365 and external data. This means you can access information from many places without leaving your main workspace. You can also use APIs to automate data flow and make your workflows more efficient.
| Role of Connectors and APIs | Description |
|---|---|
| Data Retrieval | Connectors enable the retrieval of data from external sources into Microsoft 365 Copilot. |
| Data Indexing | They index the data, making it searchable within the Microsoft 365 environment. |
| Real-time Access | Connectors allow for real-time access to external data, enhancing user experience. |
With copilot extensibility, you can create agents that use these connectors and APIs to automate tasks and provide real-time insights. This makes your work faster and more accurate.
Microsoft Graph Integration
Microsoft Graph integration gives you powerful access to your organization’s data. Copilot uses Microsoft Graph to pull information from emails, documents, Teams conversations, calendars, and SharePoint. This creates a new knowledge model for your business. You can use Business Chat to find and surface information from different apps. Knowledge flows freely, so you spend less time searching for what you need.
Copilot extensibility ensures that your data stays secure. Microsoft Graph respects your existing permissions and compliance controls. You can trust that your information is safe while you automate and streamline your workflows. By using Microsoft Graph, you can connect data from many sources and improve your enterprise ai solutions.
You can see how these integration options make Microsoft 365 Copilot a flexible and powerful tool. You can automate, connect, and personalize your work with copilot extensibility. Whether you want to create agents, link external data, or use advanced APIs, Microsoft gives you the tools to succeed.
Community Tools and Open Source
You can boost your productivity with community-driven tools and open-source projects in the microsoft 365 ecosystem. These resources help you automate tasks, manage data, and extend the power of copilot. Many developers and IT professionals use these tools every day to solve real business problems.
PnP PowerShell
PnP PowerShell gives you a set of powerful commands for managing microsoft 365 environments. You can automate site creation, user permissions, and document libraries with simple scripts. This tool works well for both IT admins and developers. You do not need to write complex code to get results. You can use PnP PowerShell to schedule tasks, migrate content, and enforce security policies. Many organizations use it to streamline operations and reduce manual work. You can also connect PnP PowerShell with copilot to trigger automation from natural language prompts. This makes your daily work faster and more reliable.
Tip: Start with basic scripts in PnP PowerShell. You can build more advanced automation as you learn.
PnP Core SDK
PnP Core SDK helps you build custom solutions for microsoft 365. You can use this software development kit to interact with SharePoint, Teams, and other services. The SDK supports both .NET and low-code approaches. You can create apps that read and write data, manage lists, and automate workflows. PnP Core SDK works well with copilot extensibility. You can build agents that use the SDK to fetch information or update records. This toolkit gives you flexibility to solve unique business challenges. You can also use it with the microsoft 365 agents toolkit to create enterprise ai solutions.
Note: PnP Core SDK is open source. You can contribute to its development or use community examples to learn new skills.
Developer Contributions
You can shape the future of microsoft 365 copilot by sharing your own solutions. Developers add value to the platform through plugins, Graph connectors, and integrations. You can use Graph connectors to bring external data into copilot. This gives your organization richer context and smarter responses. For example, Lucid’s integration lets you search and summarize content from virtual whiteboards and diagrams using natural language. You can also create plugins with Teams message extensions. These plugins help you retrieve data and take action inside Teams. Many developers share their work with the microsoft community. You can learn from their examples or join open-source projects.
- Enhance copilot with plugins and Graph connectors.
- Integrate external data sources for better insights.
- Use Teams message extensions for custom actions.
- Share your solutions with the microsoft 365 community.
You can join a global network of contributors. Your work helps others and drives innovation in microsoft 365. Open-source projects make it easier for everyone to build, automate, and personalize their workflows.
Getting Started with Custom Tools
You can start using custom tools in microsoft 365 copilot with a few simple steps. The process helps you unlock new productivity features and tailor your workspace to your needs. You do not need advanced technical skills. Many users benefit from a low-code/no-code approach, which makes setup easy and accessible.
Setup Guide
Accessing Features
You can access microsoft 365 copilot features by following these steps:
- Sign in to the Microsoft 365 Admin Center. This is where you manage subscriptions and licenses.
- Open the Billing section to view available services.
- Search for and select Microsoft Copilot based on your licensing eligibility.
- Choose your subscription type and quantity to configure your purchase.
- Review and complete your purchase to finalize the transaction.
- Assign Copilot licenses to users so your team can start using the features.
Tip: Assigning licenses early ensures your team can explore copilot extensibility and begin automating tasks right away.
Creating Tools
You can create agents and custom tools in microsoft 365 using Copilot Studio or other supported platforms. Start by selecting templates that match your workflow. Use drag-and-drop interfaces to build automation. You can personalize agents to handle tasks like document creation, scheduling, or data analysis. Many users find that customizing templates saves time and improves consistency.
Note: You can use the microsoft 365 agents toolkit to build advanced solutions or start with simple templates for quick wins.
App Integration
You can integrate microsoft 365 copilot with popular apps to streamline collaboration and data management. The microsoft 365 ecosystem supports many integration options.
Teams
You can connect copilot to Teams for enhanced meeting summaries, task assignments, and real-time collaboration. Custom agents help you automate follow-ups and organize conversations. You can use message extensions to retrieve data or trigger actions directly in Teams.
SharePoint
You can use copilot in SharePoint to automate document management and workflow approvals. Custom tools help you organize files, set permissions, and track changes. You can build agents that move documents based on rules or summarize content for easy review.
Tip: Integrating copilot with Teams and SharePoint boosts productivity and keeps your information organized.
Tips for Beginners
You can make the most of microsoft 365 copilot by starting with these actionable tips:
- Assign copilot licenses to managed users in the Microsoft 365 Admin Center.
- Select 'Users' and then 'Active users' to choose who gets access.
- Use 'Manage product licenses' to assign copilot and save changes.
- Consider device management tools like Microsoft Intune to extend copilot to mobile devices.
- Identify key business functions where copilot can help, such as finance, marketing, or sales.
- Use role-based scenario guidance to discuss workflow pain points and opportunities with your team.
Note: Start small and expand as you learn. You can build confidence by automating simple tasks before moving to more complex solutions.
You can explore copilot extensibility and create agents that fit your needs. Microsoft 365 gives you the tools to personalize your workspace and improve productivity. You can join the microsoft 365 ecosystem and discover new ways to work smarter.
Best Practices for Copilot Extensibility
Designing custom tools for microsoft 365 copilot helps you get the most from your investment. You can follow these best practices to ensure your solutions are effective, secure, and easy to maintain.
Designing Tools
Workflow Simplicity
You should keep your workflows simple and intuitive. Start by mapping out your process before building new tools. Choose solutions that integrate well with your existing microsoft systems. Involve users early in the design phase so you can address their needs and encourage adoption. Provide ongoing training and support to help your team use copilot effectively.
- Map out your process before building.
- Choose tools that fit with your current microsoft environment.
- Involve users from the start.
- Keep the user experience simple.
- Offer training and support.
Tip: Start with simple implementations and add complexity as you learn what works best.
Security
Security is critical when you develop custom tools. You should conduct regular security assessments to prevent data leaks. Use strict access controls, such as role-based permissions, to protect sensitive information. Train your team on proper data handling to meet regulations like GDPR and HIPAA. Microsoft provides built-in security features, but you must stay proactive.
- Schedule frequent security reviews.
- Set up role-based access for sensitive data.
- Train employees on safe data practices.
Maintaining Solutions
Testing
You need to test your copilot agents and custom tools thoroughly. Collect feedback from users and review performance often. Use documentation and version control to track changes. For complex projects, set up continuous integration and deployment (CI/CD) to automate testing and updates. This approach helps you catch issues early and keep your tools running smoothly.
- Maintain clear documentation.
- Use version control for all changes.
- Test regularly and gather user feedback.
- Implement CI/CD for larger projects.
Community Resources
You can use community resources to support your development and troubleshooting. Microsoft offers 24/7 technical support through the IT Service Desk. You can also join forums and share your experiences with other users. Feedback from the community helps improve copilot extensibility for everyone.
Note: Sharing your experiences and learning from others can help you solve problems faster.
Avoiding Pitfalls
You should watch out for common challenges when extending copilot. Custom development may require more effort and ongoing maintenance. You need developer expertise and the right licenses. Make sure you use a production tenant for plugin development, as developer tenants do not support this feature. Assign the correct microsoft copilot license to each user involved in development.
- Plan for higher initial effort and maintenance.
- Ensure your team has the right skills.
- Use production tenants for plugin development.
- Assign the correct licenses to developers.
By following these best practices, you can maximize the value of copilot extensibility. You will create secure, efficient, and user-friendly solutions that help your organization work smarter with microsoft 365 copilot.
You can transform your daily work with microsoft 365 copilot extensibility. Custom tools help you automate tasks, connect apps, and personalize workflows. See the benefits in the table below:
| Benefit | Example |
|---|---|
| Cost-savings | Streamlined workflows reduce manual overhead. |
| Efficiency | Single-pane views enhance collaboration. |
| Creativity | Automation lets you focus on big ideas. |
| Integration | Connect with apps like ServiceNow for IT task prioritization. |
| Accessibility | Ask copilot about benefits or PTO statistics directly. |
Start your journey with these steps:
- Compare agent results with microsoft 365 copilot.
- Verify conversation starters using step-by-step guidance.
- Build custom agents in Copilot Studio.
Ongoing learning and community engagement with microsoft copilot help you find answers faster and improve efficiency. Experiment, share feedback, and join workshops to unlock the full potential of microsoft 365.
FAQ
How do you start building custom tools in Microsoft 365 Copilot?
You can begin by accessing Copilot Studio. Choose a template or start from scratch. Use the drag-and-drop interface to create agents or workflows. Assign licenses to your team so everyone can use the new tools.
Can you integrate third-party apps with Copilot?
Yes, you can connect apps like Salesforce, SAP, or Jira using connectors and APIs. This lets you automate workflows and access data from different platforms within Microsoft 365 Copilot.
Is coding required to use Copilot extensibility?
You do not need to code for basic customizations. Copilot Studio offers low-code and no-code options. For advanced solutions, you can use APIs or SDKs, but many users succeed with templates and simple tools.
How does Copilot keep your data secure?
Copilot follows Microsoft’s security standards. It uses role-based access controls and respects your organization’s permissions. You can trust that your data stays protected while you automate and personalize workflows.
What are some examples of tasks you can automate?
You can automate:
- Document creation
- Meeting summaries
- Email drafting
- Data analysis
- Task assignments
These automations help you save time and reduce manual work.
Where can you find help or community resources?
You can visit Microsoft’s official forums, join user groups, or explore open-source projects like PnP PowerShell. Many users share tips, scripts, and templates to help you get started and solve problems.
Can you personalize Copilot for your team’s workflow?
Yes, you can create custom agents, templates, and workflows that match your team’s needs. Copilot lets you adjust tools so they fit your business processes and goals.
🚀 Want to be part of m365.fm?
Then stop just listening… and start showing up.
👉 Connect with me on LinkedIn and let’s make something happen:
- 🎙️ Be a podcast guest and share your story
- 🎧 Host your own episode (yes, seriously)
- 💡 Pitch topics the community actually wants to hear
- 🌍 Build your personal brand in the Microsoft 365 space
This isn’t just a podcast — it’s a platform for people who take action.
🔥 Most people wait. The best ones don’t.
👉 Connect with me on LinkedIn and send me a message:
"I want in"
Let’s build something awesome 👊
1
00:00:00,000 --> 00:00:05,000
Welcome to another edition of the M665 show.
2
00:00:05,000 --> 00:00:09,720
And today my guest is someone deeply embedded
3
00:00:09,720 --> 00:00:12,620
in the Microsoft 6 developer ecosystem,
4
00:00:12,620 --> 00:00:14,880
Galtan Thess, and all the people,
5
00:00:14,880 --> 00:00:17,680
you can correct me after all the people like,
6
00:00:17,680 --> 00:00:20,280
I spelled their names wrong,
7
00:00:20,280 --> 00:00:22,240
but he's a full stack developer.
8
00:00:22,240 --> 00:00:25,520
He is a five time MVP, a core contributor
9
00:00:25,520 --> 00:00:29,880
to the Microsoft 365 PMP community in house mainly.
10
00:00:30,720 --> 00:00:35,240
Mainly widely used tools like PMP PowerShell,
11
00:00:35,240 --> 00:00:39,720
PMP.net, I really interesting in the last part.
12
00:00:39,720 --> 00:00:42,880
He's also passionate about open source,
13
00:00:42,880 --> 00:00:44,740
developer community sharing knowledge
14
00:00:44,740 --> 00:00:47,520
and lately has been deep diving into the world
15
00:00:47,520 --> 00:00:49,960
of a Gendig AI, a corporate extensions,
16
00:00:49,960 --> 00:00:51,360
and all this stuff.
17
00:00:51,360 --> 00:00:56,200
So in this episode, we will explore Microsoft 365
18
00:00:56,200 --> 00:00:58,120
development in the future of co-pilot,
19
00:00:58,120 --> 00:01:01,360
open source collaborations, how AI reshading
20
00:01:01,360 --> 00:01:04,560
the way developer built products today.
21
00:01:04,560 --> 00:01:06,720
Galtan, welcome to the podcast.
22
00:01:06,720 --> 00:01:10,440
- Hey, Michal, thanks a lot for having me.
23
00:01:10,440 --> 00:01:11,680
How are you?
24
00:01:11,680 --> 00:01:14,520
- Yeah, all fine, thank you for the question.
25
00:01:14,520 --> 00:01:19,400
That's amazing, you are a five time MVP,
26
00:01:19,400 --> 00:01:24,400
and a core member at the PMP team,
27
00:01:25,480 --> 00:01:29,000
and heavily involved in the CCC's Defined
28
00:01:29,000 --> 00:01:31,120
of the Developer ecosystem.
29
00:01:31,120 --> 00:01:35,520
How would you describe your journey into the Microsoft space?
30
00:01:35,520 --> 00:01:42,480
- That's okay, it seems like a long time ago
31
00:01:42,480 --> 00:01:45,040
that I started my career.
32
00:01:45,040 --> 00:01:49,680
I started my career primarily into the old school
33
00:01:49,680 --> 00:01:53,440
SharePoint server-side development side of the things
34
00:01:53,440 --> 00:01:57,920
and gradually moved towards SharePoint online and SPFX
35
00:01:57,920 --> 00:02:02,680
and the 365 development stack.
36
00:02:02,680 --> 00:02:08,160
My journey primarily was starting as a SharePoint developer
37
00:02:08,160 --> 00:02:10,440
moving towards SharePoint consultant,
38
00:02:10,440 --> 00:02:14,560
and then also working at Valo as a Valo
39
00:02:14,560 --> 00:02:18,000
and then Starbase as a micro 365 developer,
40
00:02:18,000 --> 00:02:21,080
and currently I work at Solutions to Share.
41
00:02:21,080 --> 00:02:24,200
It's a German-Japan-based company,
42
00:02:24,200 --> 00:02:27,080
and I work here as a full stack developer focusing
43
00:02:27,080 --> 00:02:30,520
on the 365 development side of the things, so yeah.
44
00:02:30,520 --> 00:02:34,200
- Yeah, I see in your profile,
45
00:02:34,200 --> 00:02:39,200
you are really big part in the icon of open source community.
46
00:02:39,200 --> 00:02:46,280
And now we have Microsoft more than in close to system.
47
00:02:49,120 --> 00:02:54,120
What think you, or how is the contribution
48
00:02:54,120 --> 00:03:00,200
to the community especially in the open source world?
49
00:03:00,200 --> 00:03:02,840
How hallowed you this?
50
00:03:02,840 --> 00:03:07,840
Both systems or the thinking of how a company should run?
51
00:03:07,840 --> 00:03:18,680
- Yeah, so I mean, I primarily focus on SharePoint
52
00:03:19,520 --> 00:03:24,000
and the issues started around SharePoint PowerShell
53
00:03:24,000 --> 00:03:28,680
in the Microsoft 360 size world.
54
00:03:28,680 --> 00:03:33,120
So what used to happen is that Microsoft has this SharePoint
55
00:03:33,120 --> 00:03:36,560
online PowerShell, and that was a lot more focused
56
00:03:36,560 --> 00:03:40,360
on SharePoint admin side of the things.
57
00:03:40,360 --> 00:03:44,080
So one of, did I understand the question correctly?
58
00:03:44,080 --> 00:03:46,240
Will go, yeah.
59
00:03:46,240 --> 00:03:51,120
What, how did you see this both worlds a closed ecosystem
60
00:03:51,120 --> 00:03:53,240
from Microsoft and the open source world?
61
00:03:53,240 --> 00:03:57,840
- Yeah, so the issue with the share, okay.
62
00:03:57,840 --> 00:04:00,920
So what used to happen is most of the developers,
63
00:04:00,920 --> 00:04:04,320
Microsoft provides their own first party PowerShell,
64
00:04:04,320 --> 00:04:06,800
which is called SharePoint online PowerShell,
65
00:04:06,800 --> 00:04:10,360
and that is primarily focused on the SharePoint admin
66
00:04:10,360 --> 00:04:12,080
side of the things.
67
00:04:12,080 --> 00:04:17,080
So the PLP PowerShell was created by Irwin von Hunen.
68
00:04:17,080 --> 00:04:21,920
He's a multiple times Microsoft MVP,
69
00:04:21,920 --> 00:04:26,360
and he found this gap in SharePoint online,
70
00:04:26,360 --> 00:04:30,080
and then Microsoft 365 basically that if you want to do
71
00:04:30,080 --> 00:04:33,680
targeted operations against a certain SharePoint artifacts,
72
00:04:33,680 --> 00:04:38,320
such as site collection lists, files and all of those things,
73
00:04:38,320 --> 00:04:41,960
the Microsoft provided PowerShell was not super good enough.
74
00:04:41,960 --> 00:04:44,160
So, it was not super targeted for that.
75
00:04:44,160 --> 00:04:48,680
So he created that, and what I did was I started adding
76
00:04:48,680 --> 00:04:51,160
my contributions towards that and improving the code.
77
00:04:51,160 --> 00:04:56,160
So we found some gaps in the Microsoft implementation
78
00:04:56,160 --> 00:04:58,760
and the offerings that were there,
79
00:04:58,760 --> 00:05:03,160
and we decided to make this better for the whole ecosystem.
80
00:05:03,160 --> 00:05:06,000
So there are quite a few things, Microsoft is doing
81
00:05:06,000 --> 00:05:08,360
that is open source, and of course there are quite a few things
82
00:05:08,360 --> 00:05:11,360
that are super closed source, such as SharePoint PowerShell
83
00:05:11,360 --> 00:05:12,200
and all those things.
84
00:05:12,200 --> 00:05:16,960
So yeah, it's an interesting gap that we are addressing
85
00:05:16,960 --> 00:05:21,960
via PNP PowerShell, and I think open source is like the future,
86
00:05:21,960 --> 00:05:24,320
and I think they should open source
87
00:05:24,320 --> 00:05:27,600
lot more parts of the system so that people can
88
00:05:27,600 --> 00:05:31,160
contribute to it more and understand it and develop things
89
00:05:31,160 --> 00:05:32,600
around that ecosystem.
90
00:05:32,600 --> 00:05:35,720
- Yeah, I think that's really great.
91
00:05:35,720 --> 00:05:40,720
But for the people who may not know the PMP ecosystem deeply,
92
00:05:41,080 --> 00:05:46,080
what are the Microsoft C60 pattern and practice PMP PowerShell
93
00:05:46,080 --> 00:05:49,160
and especially PMP.net?
94
00:05:49,160 --> 00:05:52,960
Why are there so important for developers today?
95
00:05:52,960 --> 00:05:58,560
- Yeah, so PMP PowerShell and PMP.net,
96
00:05:58,560 --> 00:06:01,520
they both address certain gaps into the Microsoft 360,
97
00:06:01,520 --> 00:06:04,400
so if development side of the things,
98
00:06:04,400 --> 00:06:08,720
so PMP PowerShell is a collection of PowerShell,
99
00:06:08,720 --> 00:06:09,720
is a collection of PowerShell.
100
00:06:09,720 --> 00:06:12,880
And it's commonly it's that you execute against SharePoint
101
00:06:12,880 --> 00:06:17,520
and it addresses real gaps that are present
102
00:06:17,520 --> 00:06:20,320
in the Microsoft 365 ecosystem.
103
00:06:20,320 --> 00:06:23,200
So NCM goes for PMP.net as well,
104
00:06:23,200 --> 00:06:26,920
so there are quite a few PMP.net library as well,
105
00:06:26,920 --> 00:06:30,640
is it's more like an SDK and then OpenMuted SDK,
106
00:06:30,640 --> 00:06:33,640
which provides like a fluent implementation
107
00:06:33,640 --> 00:06:36,000
of certain graph endpoints.
108
00:06:36,000 --> 00:06:39,000
So both of them are quite interesting
109
00:06:39,000 --> 00:06:42,720
and they address a certain gap in the Microsoft 365 ecosystem
110
00:06:42,720 --> 00:06:45,040
for the.net stack.
111
00:06:45,040 --> 00:06:50,040
And yeah, I mean that's the PMP as in Microsoft Pattern Practice
112
00:06:50,040 --> 00:06:53,320
is it's again, it's not just for.net,
113
00:06:53,320 --> 00:06:56,160
we also have quite a few things outside of.net
114
00:06:56,160 --> 00:07:01,160
such as the M365 CLI and quite a few other things as well,
115
00:07:01,160 --> 00:07:02,160
which are there.
116
00:07:02,160 --> 00:07:04,640
So VS code extensions are there,
117
00:07:04,640 --> 00:07:08,480
quite a lot of things are there in this community driven
118
00:07:08,480 --> 00:07:09,480
initiative.
119
00:07:09,480 --> 00:07:12,680
Yeah, that's really interesting.
120
00:07:12,680 --> 00:07:16,960
You are, yeah, it's really interesting
121
00:07:16,960 --> 00:07:19,600
and you are also with your company DeepDive
122
00:07:19,600 --> 00:07:21,720
into a governance solution.
123
00:07:21,720 --> 00:07:26,720
But what excited you most about building M665 platforms right now?
124
00:07:26,720 --> 00:07:37,880
So me, I love working on co-pilot and co-pilot extensibility
125
00:07:37,880 --> 00:07:41,880
nowadays as well as using of the agent AI workflow.
126
00:07:41,880 --> 00:07:45,880
So that kind of really excites me now seeing how far the,
127
00:07:45,880 --> 00:07:50,600
you know, how far the agents have come up nowadays,
128
00:07:50,600 --> 00:07:54,240
especially remember when co-pilot first launched,
129
00:07:54,240 --> 00:07:58,000
it was like, I think it was a slightly better version
130
00:07:58,000 --> 00:08:00,880
of autocomplete or sometimes it was,
131
00:08:00,880 --> 00:08:04,480
but now it does the whole end to inside of the thing.
132
00:08:04,480 --> 00:08:07,480
So it's really interesting and that kind of excites me
133
00:08:07,480 --> 00:08:09,480
a lot, so yeah.
134
00:08:09,480 --> 00:08:12,480
Yeah, I like later a little bit DeepDive,
135
00:08:12,480 --> 00:08:15,400
especially in your experience in GitHub co-pilot.
136
00:08:15,400 --> 00:08:19,680
But so I think a little bit about,
137
00:08:19,680 --> 00:08:22,960
you have a lot of experience.
138
00:08:22,960 --> 00:08:26,360
And if someone's starting with M665,
139
00:08:26,360 --> 00:08:27,200
you will open,
140
00:08:27,200 --> 00:08:30,640
they should say focused through teams, SharePoint
141
00:08:30,640 --> 00:08:35,640
or I don't know, co-pilot or PNP, Graph API or co-pilot
142
00:08:36,880 --> 00:08:38,040
or something else.
143
00:08:38,040 --> 00:08:42,720
Wow, that's kind of a loaded question,
144
00:08:42,720 --> 00:08:47,920
but let's say nowadays a lot of focus is on co-pilot
145
00:08:47,920 --> 00:08:49,320
and co-pilot extensibility.
146
00:08:49,320 --> 00:08:53,080
So that's where a lot of excitement and a lot of investments
147
00:08:53,080 --> 00:08:55,480
are happening and that's also where a lot of things
148
00:08:55,480 --> 00:08:59,120
are being actively developed and created.
149
00:08:59,120 --> 00:09:02,640
So that is one of the things that they can focus on.
150
00:09:02,640 --> 00:09:05,800
And of course, SharePoint and Graph and T,
151
00:09:05,800 --> 00:09:08,360
and all of them, they always from the core
152
00:09:08,360 --> 00:09:10,080
of all these experiences, right?
153
00:09:10,080 --> 00:09:13,160
So co-pilot, I mean, the M3 is five co-pilot,
154
00:09:13,160 --> 00:09:16,720
extensibility and everything around that is always built
155
00:09:16,720 --> 00:09:18,560
on the underlying platform,
156
00:09:18,560 --> 00:09:23,560
which is SharePoint, Teams, Loop and Planner and all those things
157
00:09:23,560 --> 00:09:27,280
which kind of power the whole M365 ecosystem.
158
00:09:27,280 --> 00:09:30,720
So obviously the base platform is important,
159
00:09:30,720 --> 00:09:34,560
but of course it also depends on what kind of development work
160
00:09:34,560 --> 00:09:35,920
you would want to do.
161
00:09:35,920 --> 00:09:37,000
So yeah.
162
00:09:37,000 --> 00:09:42,080
- Yeah, Loop is, I have so many talks about Loop.
163
00:09:42,080 --> 00:09:44,720
It's more, from you what's the side,
164
00:09:44,720 --> 00:09:49,440
yeah, I don't know, there's more something that I'd wonder
165
00:09:49,440 --> 00:09:51,520
you can use, but no one uses it,
166
00:09:51,520 --> 00:09:56,520
but yeah, Loop is getting so important in the last few months.
167
00:09:56,520 --> 00:10:02,000
But I don't think I think a little bit about, yeah.
168
00:10:03,600 --> 00:10:06,880
About your career and your view of things.
169
00:10:06,880 --> 00:10:13,600
When we look a little bit back from today,
170
00:10:13,600 --> 00:10:16,200
like I say five years or so,
171
00:10:16,200 --> 00:10:19,000
what major shift have you seen,
172
00:10:19,000 --> 00:10:22,040
how organization approach Microsoft's,
173
00:10:22,040 --> 00:10:25,800
Microsoft 365 costibization and extability?
174
00:10:25,800 --> 00:10:32,200
- Yeah, I've seen quarter of major shift in that area.
175
00:10:32,200 --> 00:10:36,280
So, I mean, how I see here is that earlier we all had SharePoints
176
00:10:36,280 --> 00:10:39,800
server object model, then we moved towards a lot more client side.
177
00:10:39,800 --> 00:10:41,600
We have SharePoint framework,
178
00:10:41,600 --> 00:10:45,720
which is super actively used by a lot of organizations
179
00:10:45,720 --> 00:10:46,560
around the world.
180
00:10:46,560 --> 00:10:49,200
And now we also have Teams app development.
181
00:10:49,200 --> 00:10:52,360
We also have a lot of co-pilot extensibility happening
182
00:10:52,360 --> 00:10:53,760
as well around this area.
183
00:10:53,760 --> 00:10:57,800
So that has been quite a shift from SharePoint focus towards
184
00:10:57,800 --> 00:11:01,840
a lot more focus on things around Teams app,
185
00:11:01,840 --> 00:11:04,600
as well as co-pilot extensibility around that.
186
00:11:04,600 --> 00:11:06,040
And yeah, yeah.
187
00:11:06,040 --> 00:11:09,760
So, I mean, organizations are actively looking into moving
188
00:11:09,760 --> 00:11:13,600
major things, it's some SharePoint towards more focus on Teams
189
00:11:13,600 --> 00:11:14,800
or co-pilot around that.
190
00:11:14,800 --> 00:11:18,560
So, yeah, that's an interesting shift that happened here.
191
00:11:18,560 --> 00:11:20,560
- Especially in your work,
192
00:11:20,560 --> 00:11:23,840
I think one thing must be really important
193
00:11:23,840 --> 00:11:26,000
that's the Microsoft Graph API.
194
00:11:26,000 --> 00:11:30,640
What can you tell us about a Microsoft Graph API?
195
00:11:30,640 --> 00:11:35,640
And what thing will happen in the future with the Graph?
196
00:11:35,640 --> 00:11:42,720
- Yeah, Grass API, as we all know is kind of like the heart
197
00:11:42,720 --> 00:11:47,040
and soul of the ecosystem powering basically every experiences.
198
00:11:47,040 --> 00:11:52,560
So, it is also an interesting API built by different developers
199
00:11:52,560 --> 00:11:56,840
and developer teams inside Microsoft who have their own opinions
200
00:11:56,840 --> 00:11:58,480
around how to develop things.
201
00:11:58,480 --> 00:12:02,640
And some of them are not super well implemented,
202
00:12:02,640 --> 00:12:04,520
some of them are super well implemented.
203
00:12:04,520 --> 00:12:08,480
And I think it will kind of evolve.
204
00:12:08,480 --> 00:12:12,400
And now with the agentic AI happening,
205
00:12:12,400 --> 00:12:16,440
I think we should see a lot more investment, more of it,
206
00:12:16,440 --> 00:12:20,200
some graph itself and move towards agents
207
00:12:20,200 --> 00:12:24,560
and agentic API kind of things, agentic extensibility
208
00:12:24,560 --> 00:12:27,080
or those kind of things is what I think
209
00:12:27,080 --> 00:12:31,240
might have been going forward in the couple of years, I guess, yeah.
210
00:12:31,240 --> 00:12:36,760
- A little bit, before I got in these interviews,
211
00:12:36,760 --> 00:12:38,680
I scraped a little bit of profiles
212
00:12:38,680 --> 00:12:43,680
and I see you helped maintain major open source tools
213
00:12:43,680 --> 00:12:47,720
with thousands of developers worldwide.
214
00:12:47,720 --> 00:12:52,560
What do main and tame projects like PMP PowerShell
215
00:12:52,560 --> 00:12:55,320
actually look like behind the scenes?
216
00:12:55,320 --> 00:12:59,480
So, you can clear out some tips or the tech-card information.
217
00:12:59,480 --> 00:13:04,080
- Hey, I don't do anything, I just delegated to the agents nowadays.
218
00:13:04,080 --> 00:13:14,280
No, but jokes aside, yeah, it often takes a quite interesting
219
00:13:14,280 --> 00:13:17,000
effort to help maintain these libraries
220
00:13:17,000 --> 00:13:18,440
because they are providedly used
221
00:13:18,440 --> 00:13:21,320
and anything or any change that we make,
222
00:13:21,320 --> 00:13:25,320
it impacts a lot of users as well as developers
223
00:13:25,320 --> 00:13:31,120
using these SDKs and PowerShell and Prousing Engine.
224
00:13:31,120 --> 00:13:34,160
So usually what happens is we often,
225
00:13:34,160 --> 00:13:37,320
because most of our people who use the tools
226
00:13:37,320 --> 00:13:39,400
are developers themselves.
227
00:13:39,400 --> 00:13:44,400
So quite a lot of times, they often are able
228
00:13:44,400 --> 00:13:46,840
to reproduce certain issues super well.
229
00:13:46,840 --> 00:13:50,840
So that often helps us and we look into certain issues
230
00:13:50,840 --> 00:13:52,000
and we are able to fix it.
231
00:13:52,000 --> 00:13:54,840
So again, all of these are all things that we do
232
00:13:54,840 --> 00:13:57,600
outside of our active job.
233
00:13:57,600 --> 00:14:00,160
So I like working on those things,
234
00:14:00,160 --> 00:14:02,680
spending some time on those repositories,
235
00:14:02,680 --> 00:14:05,440
outside of workers, volunteering some time
236
00:14:05,440 --> 00:14:07,840
on certain issues, answering some questions
237
00:14:07,840 --> 00:14:10,960
or if there is some discussions, answer some,
238
00:14:10,960 --> 00:14:14,840
answer or put some inputs around those discussions
239
00:14:14,840 --> 00:14:15,480
or those things.
240
00:14:15,480 --> 00:14:18,160
So, yeah, it often takes a lot of time,
241
00:14:18,160 --> 00:14:20,880
but also I like spending that time, right?
242
00:14:20,880 --> 00:14:25,880
So because I like, and like a kind of a geek kind of person,
243
00:14:25,880 --> 00:14:29,120
so you know, after work, what should I do?
244
00:14:29,120 --> 00:14:32,080
Maybe spend some time looking to certain issues,
245
00:14:32,080 --> 00:14:34,840
maybe help someone, you know, figure out some things
246
00:14:34,840 --> 00:14:35,840
and those kind of things.
247
00:14:35,840 --> 00:14:36,680
So, yeah.
248
00:14:36,680 --> 00:14:39,720
- It sounds like you have the family.
249
00:14:39,720 --> 00:14:41,960
(laughs)
250
00:14:41,960 --> 00:14:42,800
Yeah.
251
00:14:44,800 --> 00:14:48,000
How do you balance community contribution, future,
252
00:14:48,000 --> 00:14:52,480
requests, governance and many tangy quality
253
00:14:52,480 --> 00:14:54,240
and your all-source projects?
254
00:14:54,240 --> 00:14:55,680
I think it's really interesting.
255
00:14:55,680 --> 00:14:58,520
- Okay, belly-gated all-to-agents.
256
00:14:58,520 --> 00:15:01,360
(laughs)
257
00:15:01,360 --> 00:15:04,880
But we would highly encourage people to contribute.
258
00:15:04,880 --> 00:15:09,360
We often receive a lot of contributions from new people,
259
00:15:09,360 --> 00:15:12,120
from time to time, it's really encouraging as well.
260
00:15:12,120 --> 00:15:15,240
So often most of them, when it comes to,
261
00:15:15,240 --> 00:15:16,760
so from my side of the things,
262
00:15:16,760 --> 00:15:18,280
it usually helped maintain power cell
263
00:15:18,280 --> 00:15:20,800
and a bit around, progeny engine
264
00:15:20,800 --> 00:15:24,040
and some around.net, PNP.net libraries.
265
00:15:24,040 --> 00:15:27,560
So usually what happens is someone raises an issue
266
00:15:27,560 --> 00:15:30,200
that they're running into in their day-to-day life.
267
00:15:30,200 --> 00:15:31,960
So most of them are developers.
268
00:15:31,960 --> 00:15:36,040
They execute certain things in their own organizations
269
00:15:36,040 --> 00:15:38,400
against certain Microsoft 365 workloads
270
00:15:38,400 --> 00:15:39,920
using the PNP tools
271
00:15:39,920 --> 00:15:41,560
and sometimes they run into some issues
272
00:15:41,560 --> 00:15:45,120
or sometimes they notice some gap that they wanted fixed
273
00:15:45,120 --> 00:15:47,640
or they, for example, they want some feature
274
00:15:47,640 --> 00:15:49,360
that they want to implement, right?
275
00:15:49,360 --> 00:15:51,840
So there is that particular request on GitHub
276
00:15:51,840 --> 00:15:56,160
and we often look into that and try to implement it
277
00:15:56,160 --> 00:15:59,760
and see how that helps them plug that gap
278
00:15:59,760 --> 00:16:02,120
or help them in their own day-to-day life.
279
00:16:02,120 --> 00:16:05,680
So usually that's how it works often.
280
00:16:05,680 --> 00:16:08,120
We also receive pull requests from certain people
281
00:16:08,120 --> 00:16:10,400
because they obviously know how to,
282
00:16:10,400 --> 00:16:12,080
the whole code is all open source.
283
00:16:12,080 --> 00:16:15,600
So they can pick into the code, see what needs to be changed
284
00:16:15,600 --> 00:16:17,120
and they raise the pull request.
285
00:16:17,120 --> 00:16:22,120
So we review them usually against our own standards
286
00:16:22,120 --> 00:16:25,680
or use some common sense around that
287
00:16:25,680 --> 00:16:28,000
and if it looks, could it not be often
288
00:16:28,000 --> 00:16:29,680
merged that pull request around that?
289
00:16:29,680 --> 00:16:30,840
So yeah.
290
00:16:30,840 --> 00:16:34,000
- Yeah, my next question is definitely,
291
00:16:34,000 --> 00:16:38,240
how have you ever been a community contribution
292
00:16:38,240 --> 00:16:41,200
or pull requests that surprised you?
293
00:16:41,200 --> 00:16:48,560
I say, surprised you in a positive or negative way
294
00:16:48,560 --> 00:16:52,800
that was where you say this was also a request?
295
00:16:52,800 --> 00:16:58,680
- Yeah, often times we often have those kind of things,
296
00:16:58,680 --> 00:17:02,320
some of, sometimes some scenarios are there
297
00:17:02,320 --> 00:17:06,520
which we maintain the code so we know quite a lot
298
00:17:06,520 --> 00:17:11,040
of things around that code, but sometimes people notice
299
00:17:11,040 --> 00:17:15,160
some really obvious gaps or some really obvious errors
300
00:17:15,160 --> 00:17:19,320
and then you realize that, oh wow, that's how they can do it.
301
00:17:19,320 --> 00:17:21,320
For example, in TNP PowerShell,
302
00:17:21,320 --> 00:17:23,400
we have a way to provision teams
303
00:17:23,400 --> 00:17:28,760
and I think someone raised a pull request around education teams
304
00:17:28,760 --> 00:17:32,280
or classroom, how to provision teams inside education tenants
305
00:17:32,280 --> 00:17:33,160
and those kind of things.
306
00:17:33,160 --> 00:17:35,560
And sometimes we even don't realize
307
00:17:35,560 --> 00:17:39,720
that this was a gap and then we're often pleasantly surprised
308
00:17:39,720 --> 00:17:43,320
like wow, that's an interesting pull request and yeah.
309
00:17:43,320 --> 00:17:48,720
- You have often say, now I think,
310
00:17:48,720 --> 00:17:53,720
I do a lot of my job and yeah, I really, I think,
311
00:17:53,720 --> 00:17:59,280
yeah, I'm also doing it nearly with a high bother,
312
00:17:59,280 --> 00:18:03,240
you especially in the way also governance and compliance
313
00:18:03,240 --> 00:18:07,360
and then I, I don't know, I use cloud code for example,
314
00:18:07,360 --> 00:18:10,720
build an app and then I ask, does it secure and say,
315
00:18:10,720 --> 00:18:15,720
oh no and then you get all these, yeah, threads,
316
00:18:15,720 --> 00:18:19,640
you have to handle and then you ask a next tool,
317
00:18:19,640 --> 00:18:21,960
I don't know, Gemini or something and says,
318
00:18:21,960 --> 00:18:23,720
oh no, it's also not secure.
319
00:18:23,720 --> 00:18:28,720
So did you think really AI is,
320
00:18:29,120 --> 00:18:34,120
could be the one thing to handle or did we need
321
00:18:34,120 --> 00:18:39,000
developers in the future, they understand the process
322
00:18:39,000 --> 00:18:42,200
and how it works and to deep dive.
323
00:18:42,200 --> 00:18:49,080
- Interesting question about our own futures.
324
00:18:49,080 --> 00:18:54,280
But yeah, I think developers will be needed
325
00:18:54,280 --> 00:18:58,640
because they might be the case that they might
326
00:18:58,640 --> 00:19:02,280
be fewer developers needed or maybe fewer people
327
00:19:02,280 --> 00:19:05,400
who might be involved in developing stuff
328
00:19:05,400 --> 00:19:09,400
but I think the base knowledge around certain languages,
329
00:19:09,400 --> 00:19:14,400
certain gut instincts often helps in certain scenarios
330
00:19:14,400 --> 00:19:18,080
because AI is really good.
331
00:19:18,080 --> 00:19:20,400
I mean, he has evolved a lot.
332
00:19:20,400 --> 00:19:25,240
It has, I mean, we've seen a lot of, at least on,
333
00:19:25,240 --> 00:19:27,440
I mean, at least on Twitter and somewhere else,
334
00:19:27,440 --> 00:19:31,120
I see a lot of people writing like million lines of code
335
00:19:31,120 --> 00:19:32,840
in like a week and all of those things, right?
336
00:19:32,840 --> 00:19:34,560
So that is really impressive,
337
00:19:34,560 --> 00:19:38,480
but also impressive would be like, is it actually working?
338
00:19:38,480 --> 00:19:41,840
There are some scenarios where AI will obviously miss
339
00:19:41,840 --> 00:19:44,840
some things and that's where the developer instinct
340
00:19:44,840 --> 00:19:46,880
or the gut knowledge should kick in
341
00:19:46,880 --> 00:19:51,640
and that's where, you know, it needs that 2% push.
342
00:19:51,640 --> 00:19:55,880
Maybe AI will be able to do like 98% of the work
343
00:19:55,880 --> 00:20:00,080
but maybe in some scenarios 2% that's where the developer
344
00:20:00,080 --> 00:20:02,640
will kick in, that's how we see.
345
00:20:02,640 --> 00:20:03,640
Yeah.
346
00:20:03,640 --> 00:20:07,120
Yeah, I prepared for this discussion then.
347
00:20:07,120 --> 00:20:10,760
There is one word I never have seen before.
348
00:20:10,760 --> 00:20:12,320
It's the PMP.NET.
349
00:20:12,320 --> 00:20:14,240
Can you a little bit tell about this?
350
00:20:14,240 --> 00:20:19,160
Yeah, PMP.NET is an SDK.
351
00:20:19,160 --> 00:20:23,440
It's a graph first SDK that is there
352
00:20:23,440 --> 00:20:26,520
and it's also fluent wrapper around the graph API.
353
00:20:26,520 --> 00:20:28,440
So it is really interesting.
354
00:20:28,440 --> 00:20:32,880
It addresses certain graphs across these different workloads
355
00:20:32,880 --> 00:20:36,720
such as SharePoint sites, Teams provisioning
356
00:20:36,720 --> 00:20:42,040
and creation and updates as well as dealing with files as well.
357
00:20:42,040 --> 00:20:44,080
And the second lead is also really interesting
358
00:20:44,080 --> 00:20:47,240
around how we handle throttling.
359
00:20:47,240 --> 00:20:51,000
So if you want to use it in large volume scenarios,
360
00:20:51,000 --> 00:20:55,400
that's how as well we really handle it well in PMP.NET.
361
00:20:55,400 --> 00:20:58,320
So if you want to make requests,
362
00:20:58,320 --> 00:21:03,120
I'll use Microsoft Graph, PMP.NET and the PMP.NET SDK
363
00:21:03,120 --> 00:21:07,120
might come into picture and help you out address some gaps.
364
00:21:07,120 --> 00:21:09,360
So that's where PMP.NET is.
365
00:21:09,360 --> 00:21:10,560
Yeah.
366
00:21:10,560 --> 00:21:12,160
This is really interesting.
367
00:21:12,160 --> 00:21:16,480
And one of the biggest advantages of using the PMP libraries
368
00:21:16,480 --> 00:21:19,640
instead of working directly with the graph API.
369
00:21:20,640 --> 00:21:25,640
So yeah, Graph API, PMP.NET is like,
370
00:21:25,640 --> 00:21:28,360
as I told you, it's kind of a wrapper around the Graph API.
371
00:21:28,360 --> 00:21:32,600
So one of the advantages is that we provide like a fluent API
372
00:21:32,600 --> 00:21:34,240
to interact with the graph graph.
373
00:21:34,240 --> 00:21:37,240
Secondly, we really handle the throttling
374
00:21:37,240 --> 00:21:40,440
and throttling scenario is really well.
375
00:21:40,440 --> 00:21:43,120
So it's graph as you know,
376
00:21:43,120 --> 00:21:45,280
if you execute a lot of requests,
377
00:21:45,280 --> 00:21:47,320
think of it like millions of requests,
378
00:21:47,320 --> 00:21:49,680
you want to make certain graphing points,
379
00:21:49,680 --> 00:21:51,280
you will obviously run into throttling
380
00:21:51,280 --> 00:21:54,280
because graph will throttle your request.
381
00:21:54,280 --> 00:21:56,920
So you need to handle those kind of scenarios.
382
00:21:56,920 --> 00:22:01,120
So that's where PMP.NET SDK and PMP PowerShell,
383
00:22:01,120 --> 00:22:03,080
both of them will come into picture.
384
00:22:03,080 --> 00:22:06,400
We handle it for you and ensure that your requests
385
00:22:06,400 --> 00:22:09,320
are gracefully handled and if you run into an actual scenario,
386
00:22:09,320 --> 00:22:11,800
we provide an error around those things
387
00:22:11,800 --> 00:22:13,920
and how you want to deal with that.
388
00:22:13,920 --> 00:22:16,200
It is also providing a lot of configuration options
389
00:22:16,200 --> 00:22:19,280
around throttling, for example, in SharePoint,
390
00:22:19,280 --> 00:22:24,040
we have this read limit and when you're reaching certain read limit,
391
00:22:24,040 --> 00:22:26,840
you know, we have a quota.
392
00:22:26,840 --> 00:22:30,360
So if you reach like 90% of the quota,
393
00:22:30,360 --> 00:22:33,440
then we start slowing down those requests and all of that.
394
00:22:33,440 --> 00:22:36,560
But if you want to slow it down around 180%
395
00:22:36,560 --> 00:22:38,960
and those kind of things, you know, you want to configure
396
00:22:38,960 --> 00:22:41,800
how you want your application to execute
397
00:22:41,800 --> 00:22:44,840
in certain scenarios, that's where PMP SDK will come in.
398
00:22:44,840 --> 00:22:46,120
- Yeah.
399
00:22:46,120 --> 00:22:50,120
- I don't know if you have seen the Wolf or Wolf Street movie,
400
00:22:50,120 --> 00:22:53,000
there is a one famous scene that's a meme,
401
00:22:53,000 --> 00:22:55,240
it's like, "Salve the spell."
402
00:22:55,240 --> 00:22:56,720
- Yeah.
403
00:22:56,720 --> 00:23:01,720
- When you sell me PMP.NET, how should I do it?
404
00:23:01,720 --> 00:23:04,560
What why should I use it?
405
00:23:04,560 --> 00:23:06,000
Why should I get it?
406
00:23:06,000 --> 00:23:09,400
Okay, it's free, but how should I get it?
407
00:23:09,400 --> 00:23:10,240
Bye.
408
00:23:10,240 --> 00:23:14,440
- Yeah, a fluent API wrapper, better throttling
409
00:23:14,440 --> 00:23:18,720
and error handling scenarios, that's how, yeah.
410
00:23:18,720 --> 00:23:20,920
Compared to the first party hardest, yeah, yeah.
411
00:23:20,920 --> 00:23:27,240
- What did you think how has the move forward
412
00:23:27,240 --> 00:23:32,480
cloud first API and API first development
413
00:23:32,480 --> 00:23:37,480
influenced the evolving of PMP tooling?
414
00:23:37,480 --> 00:23:43,720
- I think APP world,
415
00:23:43,720 --> 00:23:48,360
mainly see earlier, it used to target the three
416
00:23:48,360 --> 00:23:52,520
on-premise scenarios, three on-premise plus sharepoint online.
417
00:23:52,520 --> 00:23:56,120
So we had, we used to support SharePoint by 2013,
418
00:23:56,120 --> 00:24:00,320
2016, 2019 maybe, and I think of course, SharePoint online.
419
00:24:00,320 --> 00:24:05,320
And then nowadays, we rarely see any people,
420
00:24:05,320 --> 00:24:07,720
I mean, we have seen so many,
421
00:24:07,720 --> 00:24:10,040
so few requests around SharePoint on-prem.
422
00:24:10,040 --> 00:24:12,720
So that's how it evolved, you know,
423
00:24:12,720 --> 00:24:16,600
so like, as a lot of things are obviously all cloud first,
424
00:24:16,600 --> 00:24:19,840
they will never be delivered to the on-prem side of the things
425
00:24:19,840 --> 00:24:22,280
and the graphs, as you know, is always going to be
426
00:24:22,280 --> 00:24:24,960
on Microsoft 365, it's never going to be
427
00:24:24,960 --> 00:24:26,880
on-prem side of the things.
428
00:24:26,880 --> 00:24:31,880
So that's how the evolution evolution of the SDK has happened.
429
00:24:31,880 --> 00:24:36,640
So earlier, it used to target on-prem plus cloud
430
00:24:36,640 --> 00:24:39,880
and then it's now mostly cloud first,
431
00:24:39,880 --> 00:24:41,960
it's always going to be cloud first, now, yeah.
432
00:24:41,960 --> 00:24:42,800
- Hello.
433
00:24:42,800 --> 00:24:48,400
- Let's jump a little bit more security topic.
434
00:24:48,400 --> 00:24:51,200
How important is authentication design today
435
00:24:51,200 --> 00:24:54,480
when building enterprise-ready automation with PNP?
436
00:24:54,480 --> 00:24:58,200
- Can you come again?
437
00:24:58,200 --> 00:25:00,400
- Yeah, no problem.
438
00:25:00,400 --> 00:25:03,760
What did you think how important is authentication design today
439
00:25:03,760 --> 00:25:07,400
when build enterprise-ready automation with PNP?
440
00:25:08,680 --> 00:25:12,400
- Yeah, I think security is always the first
441
00:25:12,400 --> 00:25:14,880
and the foremost thing at the top of the mind.
442
00:25:14,880 --> 00:25:19,880
So in PNP, we support any enter ID that you have.
443
00:25:19,880 --> 00:25:24,880
So as long as, see, as same goes with the PNP.NET as well,
444
00:25:24,880 --> 00:25:29,800
as long as your app supports certain permissions,
445
00:25:29,800 --> 00:25:31,840
PNP.NET will work.
446
00:25:31,840 --> 00:25:36,360
So if there are certain scenarios where enter ID is there
447
00:25:36,360 --> 00:25:40,720
and you are only allowed to specify certain permissions,
448
00:25:40,720 --> 00:25:43,960
but PNP.NET supports a lot more things
449
00:25:43,960 --> 00:25:47,240
and in those scenarios because your enter ID
450
00:25:47,240 --> 00:25:49,720
doesn't have those permissions, the PNP will not work.
451
00:25:49,720 --> 00:25:52,400
So security is at the top of the mind
452
00:25:52,400 --> 00:25:57,400
and we always respect the enter ID permissions there.
453
00:25:57,400 --> 00:25:59,800
So I think it's at the top of the mind
454
00:25:59,800 --> 00:26:04,680
and whenever a developer has to develop on top of PNP.NET
455
00:26:04,680 --> 00:26:08,160
or PNP PowerShell on all of that, they should always consider
456
00:26:08,160 --> 00:26:10,600
the least permissions that they might need
457
00:26:10,600 --> 00:26:14,520
to execute certain operations around PNP.NET
458
00:26:14,520 --> 00:26:17,000
or PNP PowerShell or Crassey PI itself.
459
00:26:17,000 --> 00:26:18,520
So yeah, I think.
460
00:26:18,520 --> 00:26:21,080
- That's interesting.
461
00:26:21,080 --> 00:26:26,080
I think a little bit about is there misconceptions
462
00:26:26,080 --> 00:26:29,880
that you've often have about the PNP PowerShell?
463
00:26:29,880 --> 00:26:32,640
What do you expect that?
464
00:26:33,880 --> 00:26:34,920
- Yep.
465
00:26:34,920 --> 00:26:38,240
Yeah, I think one of the misconceptions is that PNP PowerShell
466
00:26:38,240 --> 00:26:40,200
is Microsoft developed.
467
00:26:40,200 --> 00:26:42,120
It is not, it is completely open source.
468
00:26:42,120 --> 00:26:45,040
It is not, it is completely open source.
469
00:26:45,040 --> 00:26:48,040
It is maintained by a team of maintainers
470
00:26:48,040 --> 00:26:51,680
who spend their time outside of work
471
00:26:51,680 --> 00:26:53,360
to help maintain these libraries
472
00:26:53,360 --> 00:26:56,720
and again, it is not supported by Microsoft.
473
00:26:56,720 --> 00:26:59,800
So whenever you reason, so oftentimes we often get
474
00:26:59,800 --> 00:27:01,640
some requests by someone that,
475
00:27:01,640 --> 00:27:06,080
the Microsoft support is telling us to use request
476
00:27:06,080 --> 00:27:08,920
or some issues around PNP PowerShell.
477
00:27:08,920 --> 00:27:11,360
And oftentimes people do raise a request
478
00:27:11,360 --> 00:27:14,040
with Microsoft support that, hey, this thing is not working
479
00:27:14,040 --> 00:27:17,240
in PNP PowerShell and then Microsoft supports,
480
00:27:17,240 --> 00:27:20,480
redirects it to us and then we have to look into that.
481
00:27:20,480 --> 00:27:23,600
So that suffers a misconception
482
00:27:23,600 --> 00:27:26,000
that PNP PowerShell is developed by Microsoft
483
00:27:26,000 --> 00:27:27,080
and supported by Microsoft.
484
00:27:27,080 --> 00:27:30,520
It is not, so it is all completely community driven
485
00:27:30,520 --> 00:27:33,920
and helped and maintained by people in their suite.
486
00:27:33,920 --> 00:27:35,000
So yeah.
487
00:27:35,000 --> 00:27:36,800
Awesome.
488
00:27:36,800 --> 00:27:41,680
We have this co-pilot,
489
00:27:41,680 --> 00:27:43,720
topic everywhere actually.
490
00:27:43,720 --> 00:27:48,000
How do you see PNP evolving along Microsoft co-pilot
491
00:27:48,000 --> 00:27:49,960
and I, I, Drift Brook flows?
492
00:27:49,960 --> 00:27:54,800
Yeah, interesting topic.
493
00:27:54,800 --> 00:27:58,360
We are actively looking into how we can integrate PowerShell,
494
00:27:58,360 --> 00:28:01,360
PNP PowerShell with the agent AI workflows.
495
00:28:01,360 --> 00:28:03,920
So we're looking into developing an NCP server
496
00:28:03,920 --> 00:28:04,760
around this.
497
00:28:04,760 --> 00:28:07,920
So whenever you want to make certain requests
498
00:28:07,920 --> 00:28:09,040
around PNP PowerShell,
499
00:28:09,040 --> 00:28:11,240
this NCP server can kick in and help you out
500
00:28:11,240 --> 00:28:14,120
in certain scenarios with certain problems.
501
00:28:14,120 --> 00:28:17,000
We're also looking into working with how we can integrate
502
00:28:17,000 --> 00:28:20,240
certain skills or develop some skills around certain topics
503
00:28:20,240 --> 00:28:21,240
in PNP PowerShell.
504
00:28:21,240 --> 00:28:24,200
So I think that's how we plan to evolve.
505
00:28:24,200 --> 00:28:28,120
And of course, PNP PowerShell is based on .NET.
506
00:28:28,120 --> 00:28:31,600
So current version is .NET 8.
507
00:28:31,600 --> 00:28:35,720
And once it gets out of support we plan to improve
508
00:28:35,720 --> 00:28:37,080
to .NET 10 as well.
509
00:28:37,080 --> 00:28:40,200
So that should bring in all the latest and greatest
510
00:28:40,200 --> 00:28:41,960
from the .NET itself,
511
00:28:41,960 --> 00:28:45,640
.NET runtime itself into the PNP PowerShell.
512
00:28:45,640 --> 00:28:50,880
And yeah, besides that PNP PowerShell works really well
513
00:28:50,880 --> 00:28:54,000
from what I understand and what I've tested in Cloud Code
514
00:28:54,000 --> 00:28:58,080
or GitHub co-pilot or Codex and all of those things
515
00:28:58,080 --> 00:29:01,000
that they do understand PowerShell,
516
00:29:01,000 --> 00:29:04,360
the PowerShell really well and they execute the commands
517
00:29:04,360 --> 00:29:07,640
and the scripts and all that because they understand the code
518
00:29:07,640 --> 00:29:09,320
and again, it's really helpful.
519
00:29:09,320 --> 00:29:11,640
Yeah.
520
00:29:11,640 --> 00:29:14,520
>> Yeah, you are currently deep dive into
521
00:29:14,520 --> 00:29:17,040
a GenTik AI development experience,
522
00:29:17,040 --> 00:29:21,160
especially because GitHub co-pilot,
523
00:29:21,160 --> 00:29:24,680
what first pulled you into the space
524
00:29:24,680 --> 00:29:28,080
or how did you major in this space?
525
00:29:28,080 --> 00:29:35,480
>> Yeah, GitHub co-pilot when it first came as it was a bit awkward,
526
00:29:35,480 --> 00:29:40,880
it was a bit okay-ish, but yeah, I think with time
527
00:29:40,880 --> 00:29:44,000
when there were new models came in and they started to start
528
00:29:44,000 --> 00:29:46,280
experimenting with them, building things
529
00:29:46,280 --> 00:29:48,720
on using those agents and all of that.
530
00:29:48,720 --> 00:29:53,120
Primarily it started, initially it was a very little use.
531
00:29:53,120 --> 00:29:55,400
Then I think GPT 3.
532
00:29:55,400 --> 00:29:58,600
Sorry, Sonnet 3.5, I think it was a 3.7,
533
00:29:58,600 --> 00:30:01,840
which was a really good model and I think one of the first models
534
00:30:01,840 --> 00:30:07,560
that kind of, it was like a bulb moment,
535
00:30:07,560 --> 00:30:11,400
you have bulb glow moment that kicked in wow,
536
00:30:11,400 --> 00:30:14,360
like this can be a start of really interesting things,
537
00:30:14,360 --> 00:30:17,920
so that's how I got pulled into that and I think
538
00:30:17,920 --> 00:30:19,720
once new models came in,
539
00:30:19,720 --> 00:30:23,720
so one of my few-rit models nowadays is GPT 5.5,
540
00:30:23,720 --> 00:30:27,400
so yeah, that model is really, really cool
541
00:30:27,400 --> 00:30:30,600
and I really use it primarily all day,
542
00:30:30,600 --> 00:30:35,400
every day for a lot of topics that we have in my current organization,
543
00:30:35,400 --> 00:30:39,720
as well as with PMP.NET and PMP PowerShell as well, yeah.
544
00:30:39,720 --> 00:30:44,600
>> Also, there was, I don't know,
545
00:30:44,600 --> 00:30:48,240
there was this moment Microsoft by GitHub
546
00:30:48,240 --> 00:30:52,320
and I think, oh, by the way, there's this stupid idea.
547
00:30:52,320 --> 00:30:56,520
And yeah, but I love GitHub, okay.
548
00:30:56,520 --> 00:31:00,160
But and I think the first co-pilot,
549
00:31:00,160 --> 00:31:02,520
I element was to GitHub co-pilot.
550
00:31:02,520 --> 00:31:08,600
But there are, I think, some misconceptions,
551
00:31:08,600 --> 00:31:12,760
what people think about the GitHub co-pilot,
552
00:31:12,760 --> 00:31:15,440
how will you explain the GitHub co-pilot?
553
00:31:15,440 --> 00:31:21,480
>> Yeah, GitHub co-pilot is an e-sendic AI platform.
554
00:31:21,480 --> 00:31:23,920
Let's put it that way because it,
555
00:31:23,920 --> 00:31:29,000
so if we, it has, it's like a collection of things that you use
556
00:31:29,000 --> 00:31:31,520
with your development workflow.
557
00:31:31,520 --> 00:31:35,960
So GitHub co-pilot has, you can use it inside VS code,
558
00:31:35,960 --> 00:31:41,040
or VS code, it is, you can all get up co-pilot also like a platform
559
00:31:41,040 --> 00:31:45,120
for different AI models, like from different AI providers,
560
00:31:45,120 --> 00:31:48,920
such as OpenEI, Anthropic, Gemini,
561
00:31:48,920 --> 00:31:53,360
and I think a few, a couple more as well.
562
00:31:53,360 --> 00:31:57,360
So you can use it in any other tools,
563
00:31:57,360 --> 00:31:59,560
such as even Open code as well.
564
00:31:59,560 --> 00:32:03,120
Secondly, it is also there on the cloud.
565
00:32:03,120 --> 00:32:07,320
So if you have a GitHub repository and a GitHub co-pilot subscription,
566
00:32:07,320 --> 00:32:10,640
you can point that agent towards certain things.
567
00:32:10,640 --> 00:32:14,560
For example, if someone raises an issue in a GitHub repository,
568
00:32:14,560 --> 00:32:17,120
you can just assign that issue to GitHub co-pilot,
569
00:32:17,120 --> 00:32:19,960
and it will basically work on that issue,
570
00:32:19,960 --> 00:32:21,200
understand the issue,
571
00:32:21,200 --> 00:32:23,240
three, three, three, all around that issue,
572
00:32:23,240 --> 00:32:26,560
and it's really interesting platform itself.
573
00:32:26,560 --> 00:32:27,880
Same goes with pull request.
574
00:32:27,880 --> 00:32:29,840
So if we have created a pull request,
575
00:32:29,840 --> 00:32:32,040
it will review that pull request itself, you know?
576
00:32:32,040 --> 00:32:35,880
So it really helps you in your own development workflow,
577
00:32:35,880 --> 00:32:39,640
and it's like a complete platform of AI tools, yeah?
578
00:32:39,640 --> 00:32:42,400
>> Well, that's really, really interesting.
579
00:32:42,400 --> 00:32:46,440
So you say it's an agente AI,
580
00:32:46,440 --> 00:32:50,800
but everybody talks about AI assistants.
581
00:32:50,800 --> 00:32:54,600
So how do you explain to the people,
582
00:32:54,600 --> 00:32:59,600
what's different between an AI assistant and an agente AI?
583
00:32:59,600 --> 00:33:05,280
>> So yeah, yeah, assistant, think of it this way,
584
00:33:05,280 --> 00:33:08,600
there is a task for you.
585
00:33:10,120 --> 00:33:14,160
It is an agente AI and AI assistant,
586
00:33:14,160 --> 00:33:16,160
think of it in those two terms.
587
00:33:16,160 --> 00:33:21,160
If there is a task, you give that task to AI,
588
00:33:21,160 --> 00:33:23,640
for example, GitHub co-pilot,
589
00:33:23,640 --> 00:33:26,040
GitHub co-pilot will understand and make some changes,
590
00:33:26,040 --> 00:33:28,160
you do it yourselves manually,
591
00:33:28,160 --> 00:33:29,680
you create the pull requests,
592
00:33:29,680 --> 00:33:31,840
and test it out all of that.
593
00:33:31,840 --> 00:33:34,440
That's kind of AI assisted workflow,
594
00:33:34,440 --> 00:33:38,840
whereas agente AI would be like the full end-to-end development
595
00:33:38,840 --> 00:33:40,520
life cycle around the agente AI.
596
00:33:40,520 --> 00:33:43,840
So you simply, there is some issue,
597
00:33:43,840 --> 00:33:45,800
you assign that to AI.
598
00:33:45,800 --> 00:33:47,680
AI will create a new branch,
599
00:33:47,680 --> 00:33:49,680
you will work on that,
600
00:33:49,680 --> 00:33:50,680
it will test it out,
601
00:33:50,680 --> 00:33:52,080
you create a pull request,
602
00:33:52,080 --> 00:33:54,200
and based on the permissions and all of that,
603
00:33:54,200 --> 00:33:55,320
you assign it,
604
00:33:55,320 --> 00:33:57,400
you might even be able to work it and test it out.
605
00:33:57,400 --> 00:33:59,040
So that's kind of like a different,
606
00:33:59,040 --> 00:34:03,600
so agente AI would be like a full end-to-end AI-driven workflow,
607
00:34:03,600 --> 00:34:06,800
and ag assistants would be like human in the loop,
608
00:34:06,800 --> 00:34:10,960
and then, human assisting humans in development, so.
609
00:34:10,960 --> 00:34:12,120
Yeah.
610
00:34:12,120 --> 00:34:14,160
That's really interesting.
611
00:34:14,160 --> 00:34:17,960
And I think a little bit about the future
612
00:34:17,960 --> 00:34:20,280
from GitHub co-pilot,
613
00:34:20,280 --> 00:34:21,520
what did you see,
614
00:34:21,520 --> 00:34:24,480
what's happened in the next few months?
615
00:34:24,480 --> 00:34:27,160
I know you are not allowed to
616
00:34:27,160 --> 00:34:30,640
probably to tell all from the Microsoft event,
617
00:34:30,640 --> 00:34:32,000
I think last month,
618
00:34:32,000 --> 00:34:35,360
so yeah, but can you tell,
619
00:34:35,360 --> 00:34:37,360
what you think,
620
00:34:37,360 --> 00:34:39,840
not what you know is from Microsoft.
621
00:34:39,840 --> 00:34:41,680
(laughs)
622
00:34:41,680 --> 00:34:46,680
Yeah, I think when you look at the industry trends as well,
623
00:34:46,680 --> 00:34:50,080
GitHub co-pilot is changing a lot,
624
00:34:50,080 --> 00:34:51,720
it's going to,
625
00:34:51,720 --> 00:34:54,520
yeah, it's from what I understand would be that,
626
00:34:54,520 --> 00:34:57,480
it's going to become a bit more expensive,
627
00:34:57,480 --> 00:34:59,200
or from what,
628
00:34:59,200 --> 00:35:02,400
think of this with the end of the subsidy regime
629
00:35:02,400 --> 00:35:03,400
that we used to,
630
00:35:03,400 --> 00:35:04,560
we're kind of used to it, right?
631
00:35:04,560 --> 00:35:06,800
So they did provide a blog post,
632
00:35:06,800 --> 00:35:09,640
they are moving away from premium requests
633
00:35:09,640 --> 00:35:12,000
to use it based kind of things.
634
00:35:12,000 --> 00:35:15,640
So yeah, it's going to be an interesting shift,
635
00:35:15,640 --> 00:35:18,720
because now we are all addicted to
636
00:35:18,720 --> 00:35:21,160
eSynthic AI and AI-driven workflows,
637
00:35:21,160 --> 00:35:24,760
and then suddenly it becomes a lot more expensive.
638
00:35:24,760 --> 00:35:26,920
So now you do really think around,
639
00:35:26,920 --> 00:35:29,160
you know, where you want to point out
640
00:35:29,160 --> 00:35:30,760
the AI to certain things.
641
00:35:30,760 --> 00:35:33,760
So yeah, I think it's going to be an interesting future,
642
00:35:33,760 --> 00:35:36,840
because yeah, yeah.
643
00:35:36,840 --> 00:35:40,080
- Was there something,
644
00:35:40,080 --> 00:35:44,080
a mind shifting between, I say, building traditional apps
645
00:35:44,080 --> 00:35:47,760
and now building AI power,
646
00:35:47,760 --> 00:35:52,680
AI-driven applications,
647
00:35:52,680 --> 00:35:54,400
what they're a change,
648
00:35:54,400 --> 00:35:56,560
or from your opinion,
649
00:35:56,560 --> 00:35:59,800
have you, or it's nearly the same, but it's faster?
650
00:36:02,880 --> 00:36:05,360
- More, I mean, I would suggest,
651
00:36:05,360 --> 00:36:08,920
I would think in this terms that it is a lot faster,
652
00:36:08,920 --> 00:36:11,400
and it is also like fail fast, you know?
653
00:36:11,400 --> 00:36:15,240
So the loop is a lot shorter for certain things.
654
00:36:15,240 --> 00:36:16,680
So if you want to develop an app,
655
00:36:16,680 --> 00:36:18,120
you just point it to AI,
656
00:36:18,120 --> 00:36:19,160
it will develop it,
657
00:36:19,160 --> 00:36:23,040
you are able to see it near real time in a few minutes,
658
00:36:23,040 --> 00:36:24,640
and then you are able to, you know,
659
00:36:24,640 --> 00:36:26,200
get it out or ship it out,
660
00:36:26,200 --> 00:36:28,080
and if you think it is good enough.
661
00:36:28,080 --> 00:36:32,840
So it allows a lot more flexibility,
662
00:36:32,840 --> 00:36:34,920
and shorter is the time,
663
00:36:34,920 --> 00:36:37,640
lines for developing of certain things, yeah.
664
00:36:37,640 --> 00:36:39,240
- Awesome.
665
00:36:39,240 --> 00:36:41,600
Yeah, I find it so interesting,
666
00:36:41,600 --> 00:36:46,680
there's happening so much from local to no code,
667
00:36:46,680 --> 00:36:48,120
or my code, I don't know,
668
00:36:48,120 --> 00:36:50,440
there are so many buzzword out there,
669
00:36:50,440 --> 00:36:52,560
sometimes I'll link a little bit,
670
00:36:52,560 --> 00:36:55,120
oh, what's happened in this world?
671
00:36:55,120 --> 00:37:01,280
But when you can give a tip to our,
672
00:37:02,240 --> 00:37:06,200
you and your developer who start his career
673
00:37:06,200 --> 00:37:08,040
with GitHub, GitHub, co-pilot,
674
00:37:08,040 --> 00:37:09,600
what will you,
675
00:37:09,600 --> 00:37:13,240
what tip can you give or what advice?
676
00:37:13,240 --> 00:37:20,920
- Understand the tool and some models well, how they work.
677
00:37:20,920 --> 00:37:24,840
It's, say it's going to be like how you enter an industry,
678
00:37:24,840 --> 00:37:28,160
you know, like you first learn how coding is,
679
00:37:28,160 --> 00:37:30,840
what coding is, you sort of learn those things,
680
00:37:30,840 --> 00:37:34,280
and then use that evolving and doing something
681
00:37:34,280 --> 00:37:36,840
like an iterative feedback loop.
682
00:37:36,840 --> 00:37:39,680
So learn the tools, for example,
683
00:37:39,680 --> 00:37:42,440
GitHub, co-pilot, or Cloud code, or codex,
684
00:37:42,440 --> 00:37:45,560
or learn how they work, secondly use certain things.
685
00:37:45,560 --> 00:37:49,440
For example, MCPs, however, or skills,
686
00:37:49,440 --> 00:37:51,080
and whatnot, that is available,
687
00:37:51,080 --> 00:37:54,040
that you can use alongside those tools.
688
00:37:54,040 --> 00:37:55,480
So that's how you can,
689
00:37:55,480 --> 00:37:58,080
equals is a developer, and of course,
690
00:37:58,080 --> 00:38:01,480
something that will never change is the fundamentals
691
00:38:01,480 --> 00:38:04,240
of certain language, if you are a tenant developer,
692
00:38:04,240 --> 00:38:07,760
well, it will always be better for,
693
00:38:07,760 --> 00:38:09,720
I mean, if you know the language,
694
00:38:09,720 --> 00:38:13,120
because you will be able to read code and understand code.
695
00:38:13,120 --> 00:38:15,240
So as a human, it is always important
696
00:38:15,240 --> 00:38:17,320
to understand code in my opinion,
697
00:38:17,320 --> 00:38:20,320
that so that you are responsible for the code
698
00:38:20,320 --> 00:38:21,720
that you are shipping,
699
00:38:21,720 --> 00:38:24,760
and you need to understand the code that you are shipping, right?
700
00:38:24,760 --> 00:38:27,280
So I mean, here will help you for sure,
701
00:38:27,280 --> 00:38:29,920
like 98% times it will help you,
702
00:38:29,920 --> 00:38:33,960
but that 2% times, yeah, if you don't know the code,
703
00:38:33,960 --> 00:38:35,280
good luck with that, so.
704
00:38:35,280 --> 00:38:40,080
- Yeah, I'm the print hello world guy.
705
00:38:40,080 --> 00:38:42,680
(laughing)
706
00:38:42,680 --> 00:38:46,600
You have, from your company side,
707
00:38:46,600 --> 00:38:50,240
you have a strong knowledge about governance,
708
00:38:50,240 --> 00:38:52,960
especially in teams and SharePoint,
709
00:38:52,960 --> 00:38:57,960
but how has governance changed in the code pilot area
710
00:38:57,960 --> 00:39:03,800
from your perspective, and how we can,
711
00:39:03,800 --> 00:39:06,880
I don't know, feel more safe, or how can company
712
00:39:06,880 --> 00:39:11,800
feel more safe because there's a lot of panic
713
00:39:11,800 --> 00:39:14,560
and a lot of other people say they can nothing,
714
00:39:14,560 --> 00:39:16,520
but what did you think?
715
00:39:16,520 --> 00:39:19,160
- Yeah, I think governance is,
716
00:39:19,160 --> 00:39:22,360
and we've always been interesting and topic,
717
00:39:22,360 --> 00:39:26,000
so mainly like how developers in the organization
718
00:39:26,000 --> 00:39:31,000
are using AI, the data in Greece, as well as the,
719
00:39:31,000 --> 00:39:33,840
I mean, how the data comes in, data comes out,
720
00:39:33,840 --> 00:39:36,200
maybe do some auditing, do some logging around,
721
00:39:36,200 --> 00:39:39,440
those side of the things, how developers are building
722
00:39:39,440 --> 00:39:42,960
those tools inside your organization, what data,
723
00:39:42,960 --> 00:39:44,560
I mean, what connections they are doing,
724
00:39:44,560 --> 00:39:47,200
for example, in low-code scenarios,
725
00:39:47,200 --> 00:39:51,280
you build some things, and that connects outside your,
726
00:39:51,280 --> 00:39:55,680
and the 365 ecosystem, maybe ensure that it is audited
727
00:39:55,680 --> 00:40:00,680
and loved and tracked somewhere, so same goes with AI,
728
00:40:00,680 --> 00:40:04,320
you need to ensure that your tools are compliant
729
00:40:04,320 --> 00:40:06,560
with your organization's policies,
730
00:40:06,560 --> 00:40:09,840
and then ensure that you also have the tools to track
731
00:40:09,840 --> 00:40:14,680
how these things are working, so that's how I see that, yeah.
732
00:40:14,680 --> 00:40:18,440
- Yeah, I am a really lazy guy,
733
00:40:18,440 --> 00:40:21,840
and I often use cloud-coder or complexity or something,
734
00:40:21,840 --> 00:40:25,880
and for simple stuff like, I don't know,
735
00:40:25,880 --> 00:40:29,720
there is a new, I don't know, for the podcast,
736
00:40:29,720 --> 00:40:34,720
there was, well, now you need also a 60 to 9 thumbnail
737
00:40:34,720 --> 00:40:38,960
to do square, there thing, and as they were,
738
00:40:38,960 --> 00:40:40,480
I'm too bored about it, they say,
739
00:40:40,480 --> 00:40:43,600
I, let us develop the net,
740
00:40:43,600 --> 00:40:46,920
and for all these things, we are in the niche,
741
00:40:46,920 --> 00:40:49,480
where everybody should build,
742
00:40:49,480 --> 00:40:52,600
like me, an app for everything,
743
00:40:52,600 --> 00:40:57,600
or is there also a space for enterprise solutions?
744
00:40:57,600 --> 00:41:05,120
- Yes, enterprise solutions will always be there,
745
00:41:05,120 --> 00:41:08,080
I think, enterprises need certain things,
746
00:41:08,080 --> 00:41:10,280
for example, reliability, availability,
747
00:41:10,280 --> 00:41:11,920
observability of some things,
748
00:41:11,920 --> 00:41:16,120
so I think that's where enterprise solutions will evolve,
749
00:41:16,120 --> 00:41:21,120
you know, maybe did I get it right,
750
00:41:21,120 --> 00:41:23,080
with you, yeah?
751
00:41:23,080 --> 00:41:26,600
- Yeah, I'm for all your side,
752
00:41:26,600 --> 00:41:28,960
it's a totally more fun question,
753
00:41:28,960 --> 00:41:32,680
but from your personal view,
754
00:41:32,680 --> 00:41:37,960
how has I changed your way for personality code,
755
00:41:37,960 --> 00:41:40,560
disease, value, or prototype solutions?
756
00:41:40,560 --> 00:41:44,560
- Yep, closing the loop faster,
757
00:41:44,560 --> 00:41:46,000
fixing the issues faster,
758
00:41:46,000 --> 00:41:50,160
making more errors, understanding those errors as well,
759
00:41:50,160 --> 00:41:52,560
better, I think oftentimes,
760
00:41:52,560 --> 00:41:54,640
there are some things I used to do earlier,
761
00:41:54,640 --> 00:41:55,960
for example, stack over a flu,
762
00:41:55,960 --> 00:41:58,640
and stack exchange and all of them.
763
00:41:58,640 --> 00:42:02,760
Now we have a living assistant inside our own
764
00:42:02,760 --> 00:42:06,480
development environment, so that has helped quite a lot,
765
00:42:06,480 --> 00:42:11,480
oftentimes, and I'm slightly a bit on the edge,
766
00:42:11,480 --> 00:42:14,680
because oftentimes, when I was a junior developer,
767
00:42:14,680 --> 00:42:17,360
I would often reach out to my seniors
768
00:42:17,360 --> 00:42:19,840
or someone in the team like,
769
00:42:19,840 --> 00:42:23,080
"Hey, help me out with this sort of things."
770
00:42:23,080 --> 00:42:27,160
Nowadays, that has shifted because everyone has AI,
771
00:42:27,160 --> 00:42:30,800
so people usually first go there,
772
00:42:30,800 --> 00:42:34,040
and then probably if they need someone,
773
00:42:34,040 --> 00:42:36,640
then they reach out to a senior developer,
774
00:42:36,640 --> 00:42:37,560
and that kind of thing,
775
00:42:37,560 --> 00:42:41,720
so that kind of evolution seems to have happened in my opinion,
776
00:42:41,720 --> 00:42:43,120
yeah?
777
00:42:43,120 --> 00:42:48,120
I think we are, all the things are so rapidly changed,
778
00:42:48,120 --> 00:42:50,440
and I think I don't know,
779
00:42:50,440 --> 00:42:54,480
every, I feel nearly every hour,
780
00:42:54,480 --> 00:42:56,040
when you look on LinkedIn,
781
00:42:56,040 --> 00:42:58,200
Microsoft has something you like,
782
00:42:58,200 --> 00:43:02,680
co-pilot notebooks,
783
00:43:02,680 --> 00:43:08,640
co-pilot skills, worker, worker IQ, and so on.
784
00:43:08,640 --> 00:43:11,760
I'm a little bit overwhelmed,
785
00:43:12,720 --> 00:43:14,480
sometimes, and I think,
786
00:43:14,480 --> 00:43:15,600
what could be interesting,
787
00:43:15,600 --> 00:43:16,840
what's not so interesting for me,
788
00:43:16,840 --> 00:43:20,160
but how do you continue learning,
789
00:43:20,160 --> 00:43:25,000
and not feel overwhelmed in such rapid-changling ecosystem?
790
00:43:25,000 --> 00:43:31,600
Yeah, well, yeah, that's a fair ask.
791
00:43:31,600 --> 00:43:34,720
I mean, it's also the thing,
792
00:43:34,720 --> 00:43:35,920
sign of things to happen,
793
00:43:35,920 --> 00:43:38,920
rate, duplication of whatever your life is,
794
00:43:40,040 --> 00:43:44,960
but, yeah, I think it kind of depends on the individual as well,
795
00:43:44,960 --> 00:43:47,920
it's like, example, if you're interested in certain topics,
796
00:43:47,920 --> 00:43:51,920
maybe just keep a track of the people who are super involved
797
00:43:51,920 --> 00:43:54,200
in those topics, follow them, maybe,
798
00:43:54,200 --> 00:43:56,720
and see certain blogs around those areas,
799
00:43:56,720 --> 00:43:58,680
follow Microsoft topics around that area,
800
00:43:58,680 --> 00:44:03,560
and I think maybe it's co-pink your attention to certain things,
801
00:44:03,560 --> 00:44:06,000
should help, because a lot of things happen,
802
00:44:06,000 --> 00:44:08,160
is a lot of areas,
803
00:44:08,160 --> 00:44:11,560
and not everything you need to know.
804
00:44:11,560 --> 00:44:14,000
For example, I don't need to know co-pilot,
805
00:44:14,000 --> 00:44:17,120
no book, I'm happy with knowing co-pilot,
806
00:44:17,120 --> 00:44:18,080
get-up co-pilot.
807
00:44:18,080 --> 00:44:21,240
So, I mean, thinking those terms, right?
808
00:44:21,240 --> 00:44:25,360
It's all, it's going to be an overview of the future, for sure.
809
00:44:25,360 --> 00:44:28,840
So, it's really important where we focus,
810
00:44:28,840 --> 00:44:30,240
for attention on, yeah.
811
00:44:30,240 --> 00:44:35,880
Okay, then we have the rapid-fire-out.
812
00:44:37,720 --> 00:44:39,400
So, I ask a question that you give us,
813
00:44:39,400 --> 00:44:44,480
short, as what do you think about your favorite developer tool right now?
814
00:44:44,480 --> 00:44:46,200
Cloud-cour.
815
00:44:46,200 --> 00:44:51,280
A one-microsoft technology, more developer-should-pay attention to it.
816
00:44:51,280 --> 00:44:53,120
Co-pilot, see your life.
817
00:44:53,120 --> 00:44:55,440
Teps or space?
818
00:44:55,440 --> 00:44:56,800
Times.
819
00:44:56,800 --> 00:45:00,760
Most other right, it's skill for developers.
820
00:45:00,760 --> 00:45:02,480
Debugging.
821
00:45:02,480 --> 00:45:05,800
One-air prediction, you stay strongly believe in.
822
00:45:06,800 --> 00:45:09,920
Yeah, you will eat software.
823
00:45:09,920 --> 00:45:13,680
Your favorite finished dish.
824
00:45:13,680 --> 00:45:18,200
Oh, wow.
825
00:45:18,200 --> 00:45:22,960
You know, a hard chocolate.
826
00:45:22,960 --> 00:45:24,480
So, be-
827
00:45:24,480 --> 00:45:26,320
Coffee-t-you.
828
00:45:26,320 --> 00:45:30,440
Coffee-t-you are energy jinx during the coding sessions.
829
00:45:30,440 --> 00:45:32,440
Water.
830
00:45:32,440 --> 00:45:33,440
Oh, right.
831
00:45:33,440 --> 00:45:37,320
Oh, yeah, it's an option I would often.
832
00:45:37,320 --> 00:45:39,520
Sorry.
833
00:45:39,520 --> 00:45:46,360
Best way to learn, documentation, videos or building projects.
834
00:45:46,360 --> 00:45:48,400
Documentation.
835
00:45:48,400 --> 00:45:52,960
A dream project you love to build someday.
836
00:45:52,960 --> 00:45:58,400
Uh, video, something.
837
00:45:58,400 --> 00:46:01,640
Video, translator, kind of thing, yeah.
838
00:46:01,640 --> 00:46:05,520
And, Rick and people follow your work and connect with your online.
839
00:46:05,520 --> 00:46:09,480
Um, LinkedIn, Twitter, X, yeah.
840
00:46:09,480 --> 00:46:17,080
So, all the links and the person profile is in the show now.
841
00:46:17,080 --> 00:46:20,600
So, a lot of people can check it out.
842
00:46:20,600 --> 00:46:27,640
So, then my last question is, the same question I have asked, what is the one thing the
843
00:46:27,640 --> 00:46:33,120
developers show take from this session today?
844
00:46:33,120 --> 00:46:40,040
Learn AI, uh, AI, AI, AI assisted development, uh, be proficient.
845
00:46:40,040 --> 00:46:41,120
That's the future.
846
00:46:41,120 --> 00:46:43,720
Uh, language fundamentals are important.
847
00:46:43,720 --> 00:46:51,000
So, pick a particular language set, for example, node.js stack or C#, ASP.NET stack or something,
848
00:46:51,000 --> 00:46:56,920
be proficient in one of them and use AI to the fullest, yeah.
849
00:46:56,920 --> 00:46:58,600
Yeah, thank you.
850
00:46:58,600 --> 00:47:01,400
So, you know, this was a great session.
851
00:47:01,400 --> 00:47:03,240
Uh, thank you so much.
852
00:47:03,240 --> 00:47:06,760
It was really interesting, especially in the PMP world.
853
00:47:06,760 --> 00:47:08,280
So, thank you so much.
854
00:47:08,280 --> 00:47:09,280
Bye.
855
00:47:09,280 --> 00:47:11,280
Thanks, we'll go thank you for having me.
856
00:47:11,280 --> 00:47:11,720
Bye.
857
00:47:11,720 --> 00:47:21,720
.

Founder of m365.fm, m365.show and m365con.net
Mirko Peters is a Microsoft 365 expert, content creator, and founder of m365.fm, a platform dedicated to sharing practical insights on modern workplace technologies. His work focuses on Microsoft 365 governance, security, collaboration, and real-world implementation strategies.
Through his podcast and written content, Mirko provides hands-on guidance for IT professionals, architects, and business leaders navigating the complexities of Microsoft 365. He is known for translating complex topics into clear, actionable advice, often highlighting common mistakes and overlooked risks in real-world environments.
With a strong emphasis on community contribution and knowledge sharing, Mirko is actively building a platform that connects experts, shares experiences, and helps organizations get the most out of their Microsoft 365 investments.

Am a full stack developer at Solutions2Share and 5x Microsoft MVP. I help maintain PnP PowerShell and PnP .NET libraries and am also a PnP Core team member. Love contributing to open source ecosystem. Currently, deep into agentic AI development lifecycle and experimenting different ways to builds products using different AI tools such as GitHub Copilot, Claude Code and Codex. Always be learning and sharing is caring is what I usually live by.









