Microsoft Foundry: Unlock Adaptive, Personalized Agents with User-Scoped Persistent Memory

2026-03-31 · Source: Microsoft Foundry Blog articles · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Software Development & Engineering · Depth: Intermediate, medium

Summary

Microsoft Foundry introduces a design pattern for user-scoped persistent memory in enterprise AI agents, addressing the challenge of maintaining conversational continuity and personalization across sessions while adhering to strict privacy, data isolation, governance, and compliance requirements. Traditional AI agents are stateless across sessions, losing context and user preferences. This solution, built on Azure Cosmos DB, isolates each user's context, storing curated, long-lived signals like preferences and summarized outcomes, rather than raw transcripts. It separates ephemeral in-session state, managed by Microsoft Foundry, from durable user memory. The architecture leverages Microsoft Entra ID for identity, Microsoft Foundry for agent runtime, MCP servers for memory and search, Azure Cosmos DB for NoSQL for persistent storage, and Azure AI Search for knowledge grounding. Key models include text-embedding-3-large, gpt-5-mini, and gpt-5.1.

Key takeaway

For AI Architects designing adaptive enterprise agents, this user-scoped persistent memory pattern offers a robust solution to achieve personalization without compromising compliance. You should consider implementing this reference architecture with Azure Cosmos DB to ensure data isolation and governance for long-term user context. This approach allows your agents to learn and adapt over time, enhancing user experience while meeting strict enterprise trust boundaries.

Key insights

User-scoped persistent memory enables adaptive enterprise AI agents while preserving data isolation and compliance.

Principles

• Separate ephemeral session context from durable user memory.
• Partition memory by user identity for isolation and governance.
• Knowledge grounding and personalized memory are complementary.

Method

Implement user-scoped persistent memory using Azure Cosmos DB partitioned by user ID, storing curated signals. Integrate with Microsoft Foundry for in-session context and Azure AI Search for knowledge grounding.

In practice

• Use Azure Cosmos DB for durable, user-isolated memory.
• Leverage Microsoft Entra ID for user authentication.
• Deploy with Azure Container Apps for scalable hosting.

Topics

• User-Scoped Persistent Memory
• Enterprise AI Agents
• Azure Cosmos DB
• Data Isolation & Governance
• Microsoft Foundry

Code references

• mardianto-msft/azure-agent-memory
• mardianto-msft/azure-agent-memory

Best for: AI Engineer, AI Architect, MLOps Engineer

Related on AIssential

• Teaching AI to Remember: Exploring Memory Store in Microsoft Foundry — Microsoft Foundry's Memory Store provides managed, user-scoped long-term memory for AI agents, enabling personalized, continuous interactions.
• Azure AI Foundry: The Architect’s Blueprint for Building Enterprise AI at Scale — Azure AI Foundry is a platform of ten capabilities requiring deliberate architectural choices and external governance layers for enterprise AI.
• Governed Memory: A Production Architecture for Multi-Agent Workflows — Governed Memory provides a shared memory and governance layer for multi-agent systems, enhancing recall and reducing token usage.
• How The Hermes Agent Memory Really Works — Hermes Agent's persistent memory uses a five-layered architecture for dynamic, autonomous AI agent operation.
• Your skills are leaving your hands. Don't let a rent-a-brain keep them. — AI agents require user-controlled, unsiloed access to personal context and work methods to be truly useful and prevent vendor lock-in.
• A Heterogeneous Temporal Memory Governance Framework for Long-Term LLM Persona Consistency — External temporal memory governance improves LLM long-term persona consistency and reduces drift.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Microsoft Foundry Blog articles.