Appearance
Agent Kit
Agent Kit helps teams turn agentic AI ideas into scoped, governed, and measurable business solutions. The kit is built for internal initiatives and client engagements where sponsors, business owners, architects, data teams, security, compliance, and operations need a shared process before committing to a pilot, production rollout, or broader adoption.
Use it to identify the right opportunities, choose the appropriate Microsoft-aligned platform path, define the required controls, and produce the artifacts needed to move from discovery to a scale decision.
The guidance is designed for Microsoft 365 Copilot, Copilot Studio, Microsoft Foundry, Power Platform, Dynamics 365, Azure, Microsoft Purview, Microsoft Defender, Microsoft Sentinel, and related Microsoft governance and observability services.
Choose What You Need
| You want to | Start here |
|---|---|
| Learn the kit through one guided pass | Get started |
| Run the work from intake to scale decision | How to run an agentic AI engagement |
| Facilitate the four workshops | How to facilitate workshops |
| Assess a request to chat with documents | How to assess a document chat use case |
| Classify, score, and route AI ideas | How to select and prioritize use cases |
| Look up lifecycle decisions and controls | Agent lifecycle roadmap |
| Fill delivery artifacts | Template index |
| Understand the operating logic | Outcome-first agent adoption |
Adoption Flow
| Stage | Workshops or reviews | Primary outputs |
|---|---|---|
| Strategy | Workshop 1: Strategy and use-case selection | Opportunity brief, outcome map, KPI baseline, use-case inventory, decision log, pilot shortlist |
| Data and architecture | Workshop 2: Data and architecture | Data readiness assessment, grounding strategy, retrieval/API/MCP decisions, platform selection, target architecture |
| Governance and risk | Workshop 3: Governance and risk | Operating model, RACI, agent registry model, governance policy set, responsible AI assessment, threat model, control mapping |
| Pilot design | Workshop 4: Pilot design | Agent charter, validation criteria, test set, red-team plan, ALM/environment strategy, rollout/change plan, observability specification |
| Implementation follow-up | Prototype, validate, operate, scale | Working prototype, validated controls, telemetry, feedback backlog, lifecycle review cadence |
Artifact Index
| Artifact | Purpose |
|---|---|
| Engagement guide | End-to-end phase-gate process and deliverables |
| Agent lifecycle roadmap | When to use Microsoft agent, automation, data, governance, build, validation, and operations capabilities |
| Workshop runbook | Agendas, participants, facilitation guidance, and outputs |
| Document chat assessment guide | Intake, corpus assessment, grounding, governance, and pilot planning for document chat requests |
| Use-case selection guide | Agent fit, prioritization, platform selection, and go/no-go logic |
| Templates | Delivery artifacts for planning, workshops, governance, validation, rollout, and operations |
| Raw research source | Exact source files preserved from .research |