Learn about delivering reliable, safe, and context‑aware data retrieval using Postgres‑backed agents. Abe Omorogbe (Microsoft) explores this in his talk “From Queries to Agents: The Next Era of Data Retrieval on PostgreSQL” at POSETTE: An Event for Postgres 2026. Abstract: As AI agents move from demos to production, the real challenge isn’t the model, it’s reliable, safe, and context‑aware data retrieval. In this talk, we explore how PostgreSQL is becoming the backbone for agent workflows through emerging retrieval patterns that begin with today’s Model Context Protocol (MCP) and point toward more unified approaches.

We’ll break down how agents interact with Postgres today using MCP servers, what goes wrong when agents generate SQL blindly, and why retrieval increasingly requires robust context correction alongside blended retrieval, vector similarity, relational SQL, and graph‑aware traversal working together to give agents a complete and reliable view of the data. 

We’ll then outline the architectural principles shaping the next generation of retrieval layers—designed to give agents controlled, high‑quality access to Postgres without bespoke glue code.

You’ll leave with a clear mental model for building Postgres‑backed agents today, and a practical roadmap for where agent retrieval at Microsoft is heading next.

Abe Omorogbe is a Senior Product Manager on Microsoft's Azure PostgreSQL AI team, where he drives the development of AI/ML, Vector Search, and GenAI capabilities for Azure Database for PostgreSQL. His work spans DiskANN (Microsoft Research's state-of-the-art vector indexing), Model Context Protocol (MCP) integrations, semantic operators, and the newly announced Azure HorizonDB. Abe works closely with cross-functional teams including Engineering, Azure OpenAI, and Azure ML Platform to deliver enterprise-grade AI capabilities.

► Video chapters:
⏩ 00:00 – Music & Introduction
⏩ 00:22 – Speaker introduction: sequel to last year's RAG talk
⏩ 01:02 – When AI agents go wrong: real horror stories
⏩ 03:08 – Where RAG falls short for complex queries
⏩ 04:29 – Agent architecture: multi-step, multi-tool retrieval
⏩ 05:28 – Model Context Protocol: usb-c for your AI
⏩ 06:46 – Failure modes: hallucination, semantics, guardrails
⏩ 08:44 – Context correction: four layers for trustworthy agents
⏩ 12:30 – Azure HorizonDB and four retrieval modes in one db
⏩ 16:54 – Demo: Blended retrieval in Product Copilot 
⏩ 20:39 – What's next: unified AI.search() in HorizonDB
⏩ 22:21 – Resources, GitHub repo, and thanks

📕 Everything you need to know about POSETTE: An Event for Postgres can be found at: https://posetteconf.com
✅ Learn more:  watch more POSETTE talks: https://aka.ms/posette-playlist

📌 Let’s connect:  
LinkedIn: https://www.linkedin.com/company/posetteconf/
X – @PosetteConf, https://x.com/PosetteConf 
Mastodon – @posetteconf, https://mastodon.social/@posetteconf
Bluesky – @posetteconf.com, https://aka.ms/posette-on-bluesky

Microsoft for Postgres LinkedIn: https://www.linkedin.com/company/microsoft-for-postgresql 

#PosetteConf #PostgreSQL #LLM

Microsoft Developer

A POSETTE 2026 talk by a Microsoft Azure Postgres PM covering the evolution from RAG to AI agents for data retrieval on PostgreSQL. The talk addresses common agent failure modes (hallucination, semantic misunderstanding, missing guardrails) and introduces a 4-layer context correction approach (schema, semantic, relationship, safety). It demos Azure HorizonDB, a new Postgres offering that bundles pgvector, DiskANN, Apache AGE (graph), full-text search, and azure_ai to enable blended multi-step retrieval via MCP. The demo shows how an agent using hybrid search, SQL filtering, graph traversal, and re-ranking produces more accurate results than plain vector RAG for complex e-commerce queries.

From Queries to Agents: The Next Era of Data Retrieval on PostgreSQL | POSETTE 2026