When should you extend Microsoft 365 Copilot rather than build a custom agent?
Why D is the answer
When Copilot's models and orchestration can meet the need
A declarative agent uses Copilot's orchestrator and models. Build custom when you need your own.
In detail
- Extending fits when Copilot's models and orchestration can meet the need.
- Build custom when you need your own models or orchestration.
Use case
An illustrative scenario, not a real incident.
A legal team wants Copilot to answer from its playbook. A declarative agent adds that knowledge and uses Copilot's own orchestrator.
See it in different ways
flowchart TD
A{"Does a prebuilt option meet the need?"} -->|Yes| B["Buy: Microsoft 365 Copilot or a prebuilt agent"]
A -->|"Mostly"| C["Extend with an agent or connector"]
A -->|No| D["Build a custom agent"]stateDiagram-v2 [*] --> Requirement Requirement --> Copilot: Copilot already meets it Requirement --> Extend: Copilot needs your data or tools Requirement --> Build: needs its own logic Build --> CustomModel: existing models fail a measured need
flowchart TD
Q{"Which models and orchestration?"}
Q -->|"Copilot's"| A0["Declarative agent"]
Q -->|"Your own"| A1["Custom agent"]Key point: Extend when Copilot's engine is enough.
Why the other options are wrong
- A. When you need a fully custom model and your own orchestration
- That calls for a custom agent.
- B. When the solution must run outside of Microsoft 365 entirely
- That calls for a custom agent.
- C. When users have no Microsoft 365 licenses at all
- Without licenses, extending Copilot does not apply.