Build context-aware agents: From data to decisions | BRK240
Amanda Silver and Marco Casalaina explain how to build context-aware AI agents by connecting them to enterprise knowledge, business data, and work signals using Foundry IQ, Fabric IQ, and Work IQ, with an emphasis on orchestration, governance, and operating within trusted boundaries.
Overview
The session focuses on designing high-performance, context-aware agents that can:
- Search across organizational knowledge
- Reason over business data
- Operate with awareness of people and work signals
- Take action within trusted boundaries through governance and orchestration
It introduces an “enterprise intelligence layer” for agents built around:
- Foundry IQ for connecting agents with enterprise knowledge
- Fabric IQ for ontology generation and semantic integration over business data
- Work IQ for incorporating people/work signals and organizational context
Session outline (from provided chapters)
Challenges in agentic AI projects due to lack of context
The speakers frame common problems in agentic AI projects when agents lack sufficient context from enterprise knowledge and data sources.
Foundry IQ: connecting agents with enterprise knowledge
The session introduces Foundry IQ as a way to connect agents to organizational knowledge so they can retrieve and use relevant context during reasoning and decision-making.
Demo: connecting a refund agent with Web IQ and Foundry IQ
A demonstration shows an example “refund agent” being connected to web/knowledge sources (Web IQ) and enterprise knowledge via Foundry IQ.
Fabric IQ: ontology generation and semantic integration
The speakers demonstrate Fabric IQ capabilities around generating an ontology and using it for semantic integration over business data.
Using an ontology as an MCP server and connecting with Foundry IQ
The session describes using the generated ontology as an MCP server and connecting it with Foundry IQ to provide structured, semantically meaningful context to agents.
Context delegation in data agents
The speakers explain the concept of context delegation for data agents—how context can be sourced and delegated from specialized layers/services rather than being embedded directly in the agent.
Work IQ: connecting to organizational knowledge and work signals
Work IQ is introduced as a way to incorporate organizational knowledge and work signals so agents can operate with awareness of people and work context.
Building secure, governed, scalable agent ecosystems
The session closes with guidance on building agent ecosystems that are:
- Secure
- Governed
- Scalable