Quick Note — Choose appropriate memory, tool, and knowledge integration services for agent solutions
AI-103 › Unit 1: Plan and manage an Azure AI solution › Choose the appropriate Foundry services for generative AI and agents › Choose appropriate memory, tool, and knowledge integration services for agent solutions
Quick Note — Choose appropriate memory, tool, and knowledge integration services for agent solutions
An agent is asked to do three things: recall a user's stated preferences months later, answer from a corpus the platform team already indexes, and keep a durable record of every conversation for audit. A team that reaches for one mechanism — usually memory — satisfies none of them properly. These are three different stores with three different jobs, and the exam rewards telling them apart.
Retrieval target
- Choose appropriate memory, tool, and knowledge integration services for agent solutions
- Closed-book recall
- 5 minutes
- Open Microsoft Learn after a miss
Decision anchors
| Prompt | Compact answer |
|---|---|
| Memory tool | Durable recall of facts about the user across sessions. In preview — no SLA, not recommended for production. Not a grounding corpus and not an audit record. |
| File Search | Grounds an agent in files uploaded to it, using vector search. Right for documents a user hands the agent; wrong for a governed enterprise corpus that already has a pipeline. |
| Azure AI Search tool | Grounds agents "with data from an existing Azure AI Search index" — no re-ingestion, and the owning team's enrichment and relevance tuning keep applying. The answer whenever a stem says an index is already maintained. |
| Foundry IQ knowledge base | One or more knowledge sources, an optional LLM for planning and synthesis, and retrieval parameters. Use it when several collections must be planned and merged by the service, and when grounding must be permission-aware. |
| Conversation state | The record of what was said. Use bring your own resources — your own Azure Cosmos DB — when retention and control must sit with your organization. store=false is the different requirement: do not persist at all. |
| Toolbox | A curated set of tools defined once and exposed through a single MCP-compatible endpoint, consumable by any MCP-compatible runtime. The recommended way to give agents tools. Versioned: create a version, test it, promote it to default. |
| Which tool type | Built-in — the service executes (web search, code interpreter, file search, Azure AI Search). OpenAPI tool — an external HTTP API with a specification. Function calling — your application executes and returns the result. MCP — tools shared across agents or owned by another team. |
Read the answers once, then cover the right-hand column and reconstruct each one from the prompt. A useful answer names the requirement, the capability that satisfies it, and the nearest alternative it rejects. If you can only recognize the answer after seeing it, retrieval is not yet secure.
Ninety-second explanation
Without notes, explain:
- Distinguish memory from conversation state in one sentence each.
- A corpus is already indexed and tuned by another team — what do you attach, and why not File Search?
- What preview status must a readiness review record, and what follows from it?
- What does toolbox versioning let you do that editing a tool in place does not?
- Which knowledge-source kind supports per-user permission inheritance?
Then check yourself against current Microsoft Learn. Record the missing decision rule, not merely the missed product name, in your error log.
When to go deeper
Go deeper when the design depends on agent identity and authentication for a tool — key-based, Microsoft Entra via managed identity, or OAuth On-Behalf-Of passthrough, which is what makes retrieval respect the calling user's own permissions.
Source and freshness
Written against current Microsoft Learn documentation for the AI-103 skills measured (16 April 2026). Reviewed 2026-08-20. Microsoft Learn controls changing product contracts — availability, limits, preview status, naming, and retirement dates move, and this note is deliberately compact.