Turn Meeting Transcripts into Project Plans with GitHub Copilot Agents | MVP Unplugged
Mahmoud A. Atallah walks through an “Azure Delivery Factory” approach for converting unstructured customer inputs (meeting transcripts, proposals, PDFs, briefs) into consistent, reusable delivery artifacts using GitHub Copilot agents in VS Code.
Overview
In this MVP Unplugged episode, Mahmoud A. Atallah explains and demonstrates a custom-agent workflow that:
- Converts source materials (transcripts, proposals, PDFs, briefs) into structured Markdown
- Extracts scope and requirements
- Uses GitHub Copilot custom agents, skills, and scripts to generate delivery artifacts (project plans, runbooks, task lists, risks, kickoff questions, executive summaries)
- Enriches outputs with up-to-date Azure guidance using Azure MCP Server and Microsoft Learn MCP Server, aligned to the Azure Well-Architected Framework
- Improves repeatability and trust through reusable automation (JSON, Python, PowerShell), model selection, command approvals, and human validation
What the workflow produces
Mahmoud’s agent-driven process is aimed at producing “execution-ready” artifacts such as:
- Structured requirements derived from raw customer inputs
- Project plans
- Runbooks
- Task lists (including owners)
- Risks and risk registers
- Kickoff questions
- Executive summaries
- Reusable delivery artifacts that can be applied across engagements
GitHub Copilot agent approach (custom agents, skills, scripts)
Mahmoud describes using GitHub Copilot in VS Code to create repeatable delivery workflows by combining:
- Custom agents to drive end-to-end transformations from raw inputs to deliverables
- Skills to encapsulate repeatable steps and domain-specific behaviors
- Scripts and automation to standardize outputs and reduce manual effort
Key ideas highlighted:
- Start by normalizing inputs into structured Markdown so downstream steps are consistent.
- Use repository context and existing artifacts so the agent can update and extend work rather than starting from scratch.
Adding Azure guidance with MCP servers
The episode highlights enriching agent outputs with Microsoft guidance using:
- Azure MCP Server
- Microsoft Learn MCP Server
The goal is to pull in current, authoritative references and align recommendations with the Azure Well-Architected Framework.
Controlling cost, speed, and reliability
Mahmoud calls out several practical controls for keeping AI-assisted delivery predictable:
- Reusable JSON, Python, and PowerShell automation to reduce LLM token usage and processing time
- Thoughtful model selection depending on the task
- Command approvals for safer execution
- Human validation to improve trust and catch errors before outputs become deliverables
Multi-agent orchestration
Mahmoud also discusses how multi-agent orchestration could extend the same delivery automation patterns beyond project delivery into areas like:
- Community
- Events
- Marketing
- Communications
Chapter markers (from the video)
- 00:00 – Intro to MVP Unplugged
- 00:20 – Meet Microsoft MVP Mahmoud A. Atallah
- 01:13 – Azure Delivery Factory: From Meeting Transcripts to Project Plans
- 04:16 – GitHub Copilot Custom Agents, Skills, and MCP Servers
- 08:47 – Turning Customer Calls into Structured Requirements
- 13:20 – Demo: Discovery Transcript to Delivery Artifacts
- 17:36 – Demo: Technical Proposal PDF to Project Plan
- 20:19 – JSON, Python, and PowerShell Automation for Faster Agents
- 23:57 – Reviewing Project Plans, Tasks, Risks, and Owners
- 27:47 – Multi-Agent Orchestration and Mahmoud’s MVP Journey
Links and resources
- Mahmoud's AI Agent Skills repository: https://github.com/3tallah/awesome-ai-agent-skills
- Azure Well-Architected Framework: https://learn.microsoft.com/azure/well-architected
- Azure MCP Server: https://learn.microsoft.com/azure/developer/azure-mcp-server/
- Microsoft Learn MCP Server: https://learn.microsoft.com/training/support/mcp
- GitHub Copilot: https://github.com/features/copilot
- Mahmoud on LinkedIn: https://www.linkedin.com/in/mahmoudatallah
- Justin Garrett on LinkedIn: https://www.linkedin.com/in/justgar/