--- title: Why R&D Data Belongs in the Lakehouse - and Why Agents... description: cellcentric, a hydrogen fuel cell JV between Daimler Truck and Volvo Group, shares how four years of lakehouse investment on Azure and Databricks became the... canonical: https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Why R&D Data Belongs in the Lakehouse - and Why Agents Need It There | daily.dev og:description: cellcentric, a hydrogen fuel cell JV between Daimler Truck and Volvo Group, shares how four years of lakehouse investment on Azure and Databricks became the... og:url: https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj og:image: https://api.daily.dev/og/posts/3wfyqsyFj.png og:image:alt: Why R&D Data Belongs in the Lakehouse - and Why Agents Need It There og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Why R&D Data Belongs in the Lakehouse - and Why Agents Need It There **[databricks](https://daily.dev/sources/databricks)** · 11 min read · 1 upvotes · 0 comments ## Summary cellcentric, a hydrogen fuel cell JV between Daimler Truck and Volvo Group, shares how four years of lakehouse investment on Azure and Databricks became the foundation for governed industrial AI. The core pattern: data products with rich markdown documentation, Unity Catalog for governance, and an MCP server that exposes the same governed context to AI agents as to human users. Key elements include the Fuel Cell Passport (a data product integrating SAP S/4HANA, two MES systems, a lab database, and IoT telemetry), a context-coverage quality metric that drove 90% column-comment coverage across 27 published products, and an identity model using OAuth 2.0 on-behalf-of token exchange so agents cannot bypass Unity Catalog permissions. MLflow tracing and a continuous evaluation framework monitor agent behavior, tool calls, and context drift. The result is a single governed entry point for employees and agents alike, with engineering workflows now using AI assistants connected to the same context-layer MCP during development — cutting multi-week investigations to days. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does Databricks Unity Catalog control what an AI agent can access compared to a human user? Agents access data through the same permission boundary as the human user on whose behalf they act, using OAuth 2.0 on-behalf-of token exchange starting from an identity provider like Azure AD and flowing through Databricks. Unity Catalog makes the final authorization decision, so if a user lacks access to a table or masked column, the agent hitting that same data receives the identical restriction rather than a separate, looser credential path. _Anyone designing agent access controls around Unity Catalog can find real-world governance patterns like this on daily.dev._ ### What is a context-coverage badge and why would a data team track it as a quality metric? A context-coverage badge shows whether table descriptions and column-level documentation exist across a data product's attached tables, making documentation gaps visible directly where the data is consumed. It is not proof that ambiguity is resolved, but a practical proxy that turns documentation from an after-the-fact cleanup task into something visible and improvable during delivery; one team reported 90% average column-comment coverage across 27 published products after introducing it. _Teams rethinking data quality metrics for AI-ready pipelines can track approaches like this on daily.dev._ ### Why do AI agents need more than schema metadata to reason reliably over enterprise data? Agents need to understand which business questions a data product supports, how its tables relate, what caveats apply, and which related assets should be used alongside it, not just column names and types. This richer context is captured in markdown catalog entries written during development while project context is fresh, then persisted as structured product context that both humans and agents can reason over later. _Developers wiring agents into real data platforms can follow context-layer design discussions like this on daily.dev._ ## Similar posts on daily.dev - [Why AI agents need an AI lakehouse in the modern enterprise](https://daily.dev/posts/why-ai-agents-need-an-ai-lakehouse-in-the-modern-enterprise-xux8icgxw) · Dynatrace · 3 upvotes · 0 comments - [From Legacy to Lakehouse: How Mazda Accelerated GenAI for Technical Service Operations](https://daily.dev/posts/from-legacy-to-lakehouse-how-mazda-accelerated-genai-for-technical-service-operations-xpusltb30) · databricks · 0 upvotes · 0 comments - [Unifying Data and Governance in the Agentic Era: What’s New with Azure Databricks](https://daily.dev/posts/unifying-data-and-governance-in-the-agentic-era-what-s-new-with-azure-databricks-hblz5jw8h) · databricks · 1 upvotes · 0 comments - [Why AI agents need a Context Lake](https://daily.dev/posts/why-ai-agents-need-a-context-lake-bdkjprhe2) · The New Stack · 1 upvotes · 0 comments - [Rebuilding the data stack for AI](https://daily.dev/posts/rebuilding-the-data-stack-for-ai-8s5vdavcw) · MIT Technology Review · 1 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#mcp](https://daily.dev/tags/mcp), [#data-engineering](https://daily.dev/tags/data-engineering), [#databricks](https://daily.dev/tags/databricks), [#unity-catalog](https://daily.dev/tags/unity-catalog) [View this post on daily.dev](https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Why R&D Data Belongs in the Lakehouse - and Why Agents Need It There","url":"https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj"},"datePublished":"2026-07-21T19:44:07.462Z","dateModified":"2026-09-14T08:21:03.923Z","description":"cellcentric, a hydrogen fuel cell JV between Daimler Truck and Volvo Group, shares how four years of lakehouse investment on Azure and Databricks became the...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/39f6102adda27768cf453b24a5888548?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/39f6102adda27768cf453b24a5888548?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"databricks","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"databricks","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/fa7aa720f2db4d1eba826814730482c8","url":"https://daily.dev/sources/databricks"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,mcp,data-engineering,databricks,unity-catalog","timeRequired":"PT11M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"databricks","item":"https://daily.dev/sources/databricks"},{"@type":"ListItem","position":3,"name":"Why R&D Data Belongs in the Lakehouse - and Why Agents Need It There"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/why-r-d-data-belongs-in-the-lakehouse---and-why-agents-need-it-there-3wfyqsyfj#faq","mainEntity":[{"@type":"Question","name":"How does Databricks Unity Catalog control what an AI agent can access compared to a human user?","acceptedAnswer":{"@type":"Answer","text":"Agents access data through the same permission boundary as the human user on whose behalf they act, using OAuth 2.0 on-behalf-of token exchange starting from an identity provider like Azure AD and flowing through Databricks. Unity Catalog makes the final authorization decision, so if a user lacks access to a table or masked column, the agent hitting that same data receives the identical restriction rather than a separate, looser credential path. Anyone designing agent access controls around Unity Catalog can find real-world governance patterns like this on daily.dev."}},{"@type":"Question","name":"What is a context-coverage badge and why would a data team track it as a quality metric?","acceptedAnswer":{"@type":"Answer","text":"A context-coverage badge shows whether table descriptions and column-level documentation exist across a data product's attached tables, making documentation gaps visible directly where the data is consumed. It is not proof that ambiguity is resolved, but a practical proxy that turns documentation from an after-the-fact cleanup task into something visible and improvable during delivery; one team reported 90% average column-comment coverage across 27 published products after introducing it. Teams rethinking data quality metrics for AI-ready pipelines can track approaches like this on daily.dev."}},{"@type":"Question","name":"Why do AI agents need more than schema metadata to reason reliably over enterprise data?","acceptedAnswer":{"@type":"Answer","text":"Agents need to understand which business questions a data product supports, how its tables relate, what caveats apply, and which related assets should be used alongside it, not just column names and types. This richer context is captured in markdown catalog entries written during development while project context is fresh, then persisted as structured product context that both humans and agents can reason over later. Developers wiring agents into real data platforms can follow context-layer design discussions like this on daily.dev."}}]} ```