Starter kits for agentic AI teams
Browse reusable blueprints, accelerators, and architecture samples inspired by the production patterns we use on Azure every week.
Azure AI Foundry Landing Zone Starter
A ready-to-adapt Azure foundation for secure AI experimentation with network boundaries, managed identity, and environment separation built in.
Best suited for
Cloud platform teams and solution architects
Core stack
What's included
- Opinionated resource layout for dev, test, and production environments
- Identity and access control checklist with least-privilege defaults
- Logging, alerting, and cost-observability starter configuration
Expected outcomes
- Faster tenant onboarding for AI pilots
- Lower architecture review overhead
- Governance controls in place from day one
Semantic Kernel Multi-Agent Orchestrator
A modular orchestration pattern for planner, worker, and reviewer agents with structured outputs and resilient retries.
Best suited for
Engineering teams building coordinated agent workflows
Core stack
What's included
- Typed plugin contracts and JSON-based response schemas
- Supervisor/worker pattern with handoff and retry guidance
- Evaluation hooks for task quality, latency, and cost tracking
Expected outcomes
- Reduced orchestration boilerplate
- More reliable tool-calling flows
- Easier extension as new agent skills are added
Foundry IQ Knowledge Assistant Blueprint
A retrieval-ready assistant pattern for SharePoint, OneLake, and curated data sources with citation handling and access-aware responses.
Best suited for
Teams connecting enterprise knowledge to copilots
Core stack
What's included
- Source-to-answer flow mapped for governance-heavy environments
- Prompt patterns for grounded responses, citations, and escalation
- Deployment notes for web, Teams, and internal portal experiences
Expected outcomes
- Faster time to a governed knowledge assistant
- Stronger answer traceability
- Reusable retrieval patterns across business units
Compliance Review Copilot Sample
A regulated-workflow sample that demonstrates document review, policy checks, exception routing, and human approval loops.
Best suited for
Financial services, healthcare, and operations teams
Core stack
What's included
- Reference workflow for intake, review, and approval orchestration
- Escalation and audit logging pattern for regulated changes
- Scorecards for throughput, quality, and compliance exceptions
Expected outcomes
- Clear pattern for human-in-the-loop governance
- Reusable audit trail structure
- Faster stakeholder sign-off on regulated agent workflows
Customer Service Agent Console
A service console concept showing intent routing, escalation logic, and agent assist guidance across support channels.
Best suited for
Operations teams designing omnichannel support
Core stack
What's included
- Intent-routing patterns for chat, email, and service handoff
- Agent assist views for summaries, next-best actions, and escalation notes
- KPI layer for resolution speed, automation rate, and CSAT trends
Expected outcomes
- Reusable UX for support copilots
- Shared metrics model for service leaders
- Better fit between AI assist and human escalation
Evaluation & Guardrails Toolkit
A lightweight toolkit for scorecards, regression checks, and deployment guardrails before changes reach production.
Best suited for
Teams operationalizing prompt, tool, and workflow quality
Core stack
What's included
- Example evaluation rubric for factuality, policy adherence, and completeness
- Regression test harness for prompt and tool behavior changes
- Release checklist covering rollout, rollback, and monitoring readiness
Expected outcomes
- Lower risk when shipping prompt updates
- Faster review cycles for AI changes
- Clearer quality gates for production deployments