# How a data scientist at Kilo built the business context layer his AI agents were missing

DevFeed: [How a data scientist at Kilo built the business context layer his AI agents were missing](<https://devfeed.tech/articles/how-a-data-scientist-at-kilo-built-the-business-context-layer-his-ai-agents-were-missing-101822.md>)

Original publisher: [Read original article](<https://www.getdbt.com/blog/business-context-layer-ai-agents-dbt-mcp>)

Author: Daniel Poppy

Published: 2026-10-26T14:03:00Z

Content type: article

Language: en

Sources: [dbt Labs](<https://devfeed.tech/sources/dbt-labs.md>)

Topics: [knowledge-management](<https://devfeed.tech/topics/knowledge-management.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [context](<https://devfeed.tech/topics/context.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-context](<https://devfeed.tech/tags/ai-context.md>), [ai-gateway](<https://devfeed.tech/tags/ai-gateway.md>), [anaconda](<https://devfeed.tech/tags/anaconda.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-engineering](<https://devfeed.tech/tags/analytics-engineering.md>), [atomic](<https://devfeed.tech/tags/atomic.md>), [building](<https://devfeed.tech/tags/building.md>), [business](<https://devfeed.tech/tags/business.md>), [business-context](<https://devfeed.tech/tags/business-context.md>), [cli](<https://devfeed.tech/tags/cli.md>), [context](<https://devfeed.tech/tags/context.md>), [context-layer](<https://devfeed.tech/tags/context-layer.md>), [data](<https://devfeed.tech/tags/data.md>), [data-scientist](<https://devfeed.tech/tags/data-scientist.md>), [dbt](<https://devfeed.tech/tags/dbt.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [learn](<https://devfeed.tech/tags/learn.md>), [organizational](<https://devfeed.tech/tags/organizational.md>)

## AI overview

Heyerdahl argues that a well-documented dbt project and tools such as dbt MCP, dbt skills, Wizard CLI, and dbt Semantic Layer can help agents understand models and metrics, but cannot supply the organizational history needed to interpret business decisions. A single memory file quickly went stale; raw-material search and direct Slack or Jira access still left agents to choose sources and resolve conflicts. His pipeline evolved from summarized articles, which could lose details, to atomic facts, decisions, entities, and prescriptions, then added temporal validity and supersession so historical decisions remain available without appearing current. The system uses cheaper models for routine extraction and more capable models for judgment-heavy steps. A reflect stage flags conflicts or stale information for human review, and agent findings can be submitted for approval. The excerpt presents the workflow but does not establish its measured performance.

## Source excerpt

How a data scientist built an AI context layer for organizational knowledge a dbt project alone can't hold.