Plan and manage an Azure AI solution 1.1 Choose the appropriate Foundry services for generative AI and agents

A travel agent built with Foundry Agent Service memory (preview) must use each traveler's…

Study guide skill: Choose appropriate memory, tool, and knowledge integration services for agent solutions

Question B8-003

A travel agent built with Foundry Agent Service memory (preview) must use each traveler's language preference from the start of every conversation, and should also bring up what was discussed in earlier conversations when it becomes relevant during a session. Which two statements match the documented memory types? (Choose two.)

Select 2 answers.

  • A Chat summary memory holds distilled summaries of prior conversations and is retrieved per turn by using the current messages ✓ Correct answer
  • B User profile memory holds durable preferences such as language and is retrieved near the beginning of each conversation ✓ Correct answer
  • C Store the preferences in a vector store and attach the file search tool
  • D Foundry memory is short-term memory that tracks only the current session
  • E Procedural memory is the type that stores language preferences

Answer: A, B

User profile memory stores durable preferences such as language and is retrieved near the beginning of each conversation, while chat summary memory stores summaries of prior conversations and is retrieved per turn for continuity. Procedural memory holds reusable how-to routines. Foundry memory is designed for long-term memory across sessions, and file search is for user-provided documents during an interaction.

Deep dive

How it works

Memory in Foundry Agent Service (preview) is a managed, long-term memory solution. It extracts and stores three types of long-term memory, and Microsoft gives each one its own retrieval guidance:

  • User profile memory: durable user preferences and personal context, such as language preference, product defaults or accessibility needs. Retrieve it near the beginning of each conversation to set stable personalization context.
  • Chat summary memory: distilled summaries of prior conversation topics and threads. Retrieve it per turn, using the current conversation messages, to surface relevant continuity context.
  • Procedural memory: reusable how-to routines and operating patterns inferred from prior interactions. Retrieve it when the user asks for a recurring workflow or task the agent has handled before.

The travel agent's two needs map to the first two types: the language preference is a user profile memory loaded at the start, and "what we discussed last time" is a chat summary memory pulled in when the current messages make it relevant.

Memory also needs protecting: Microsoft recommends guarding it against threats such as prompt injection and memory corruption.

The other options, and when you'd use them

  • Vector store + file search: File search is for searching user-provided documents during an interaction. Learned, per-user preferences are what memory is for.
  • Foundry memory is short-term: It's designed for long-term memory, available across sessions.
  • Procedural memory stores language preferences: Procedural memory holds reusable how-to routines. Preferences such as language belong to user profile memory.

Exam tip

Preferences (language, defaults) = user profile, loaded at the start. Past conversations = chat summary, per turn. Routines = procedural.

References

Last reviewed:

Ready to test yourself? The quiz asks this and every other AI-103 question, with instant feedback. Practice this in a quiz →

All Plan and manage questions · AI-103 study guide

Thanks for your feedback!