On-Demand # Collate Summit '26: Data & AI in **Production** Every keynote, customer story, panel, and product demo from Collate Summit '26. Watch on your schedule. [Watch the sessions](#keynotes) OpenAI, Collate leadership ## Keynotes Opening Keynote ### Why AI for Data Needs a New Foundation ![Collate](/images/collate-logo.svg) Suresh Srinivas CEO & Co-Founder, Collate Suresh opens with the question data leaders are quietly asking: why does AI for coding work, but AI for data keep falling short? The answer is not the model. It is the open context layer that grounds it. Keynote ### Inside OpenAI's Internal AI Data Agent ![OpenAI](/images/summit-recap/openai-dark.svg) Bonnie Xu Member of Technical Staff, OpenAI OpenAI's data platform serves 80% of the company. Bonnie walks through Kepler, the internal AI data agent built on OpenMetadata that takes any employee from question to insight in minutes — and the six layers of context that make it work. Product Keynote ### Rethinking How Data Works: A Collate 2.0 Deep Dive ![Collate](/images/collate-logo.svg) Sriharsha Chintalapani CTO & Co-Founder, Collate What used to take months of ticket queues can now happen in a single chat. Harsha walks through how Collate 2.0 rethinks data operations — AI agents that automate governance and quality, a Context Center that unifies knowledge, and automations that bring data to the people driving decisions. Scout24, Yelp, Ambry Genetics, Rakuten, Unity, UnionBank ## Customer Stories Lightning Talk ### AI Powered Context Catalog: Turning Institutional Knowledge into Trusted AI Context ![Scout24](/images/summit/scout24.svg) Angelita Frozza Sanches Head of Data Platform Engineering, Scout24 Scout24 moved from a legacy data catalog to a "context catalog" where AI enriches metadata, metadata powers AI agents, and governance travels with the data. Angelita shares the AI Virtuous Cycle and the confidence-scoring model behind every agent response. Lightning Talk ### Brick by Brick: Data Foundations in an Agentic World ![Yelp](/images/summit-recap/yelp-color.svg) Amy Forrest Software Engineer, Core Data Platform, Yelp Yelp switched from a duplicate-riddled legacy catalog to OpenMetadata at scale — 100k+ entities, persona-aware policies, custom search algorithms, and a read-only MCP tuned for how LLMs search. Their optimizations are now upstream contributions to OpenMetadata. Lightning Talk · Track 1 ### Clinical Data Products and Data Governance ![Ambry Genetics](/images/summit-recap/ambry-genetics-dark.png) Dan Kostecki Data Engineer II, Ambry Genetics Thirty years of hereditary cancer data, two clinical organizations, one governance platform. Dan walks through how Ambry unified data models through the Tempus acquisition and stood up a PHI-free research environment with 11,000+ assets under CAP, CLIA, HIPAA, and FDA. Lightning Talk · Track 2 ### OpenMetadata Journey at Rakuten: Unifying Assets, Enabling Discovery ![Rakuten](/images/summit-recap/rakuten-dark.svg) Muqtafi Akhmad Assistant Manager, Data Pipeline Platform Team, Rakuten Rakuten Group spans 70+ services across 30+ countries, with catalogs fragmented across spreadsheets and custom apps. Muqtafi shares the consolidation onto OpenMetadata and the path from 2–3 month data discovery to under an hour. Lightning Talk · Track 1 ### Metadata at the Moment of Truth: Compliance Reviews and Incident Triage with Collate MCP ![Unity](/images/summit-recap/unity-dark.svg) Jeppe Johansen Senior Manager, Data Platform, Unity Unity built Dax, a Slack bot that uses Collate MCP to answer compliance review and incident triage questions in 7 minutes — schema checks, PII tagging, glossary coverage, and blast-radius queries — all from the catalog as infrastructure. Lightning Talk · Track 2 ### From Manual to Observable Governance: UnionBank Philippines ![UnionBank of the Philippines](/images/summit-recap/union-bank-dark.svg) Cirene Simon Simbahan Head, Data Trust and Governance, UnionBank of the Philippines UnionBank Philippines moved from spreadsheets and tribal knowledge to observable governance across 40,000+ data assets — with a four-tier data grading policy, average issue-resolution SLAs cut from 5 days to 2, and the new Data Trust and Governance Office leading the shift. Airbus, Scout24, Ambry Genetics ## Panel: Governance in an AI-First World Panel ### Governance in an AI-First World ![Scout24](/images/summit/scout24.svg) Angelita Frozza Sanches Head of Data Platform Engineering, Scout24 ![Airbus](/images/brands/Airbus.png) Lukas Patzke Analytics Architect, Airbus ![Ambry Genetics](/images/summit-recap/ambry-genetics-dark.png) Dan Kostecki Data Engineer II, Ambry Genetics Three practitioners from regulated, enterprise, and high-scale environments on what governance has to do when AI agents are the new primary consumer of data. Moderated by Steve Wooledge, Head of Marketing at Collate. Collate 2.0 in action ## Product Demos Product Demo · Track 1 ### Creating Context: Building the AI-Ready Foundation ![Collate](/images/collate-logo.svg) Mohit Yadav Domain Lead, Collate ![Collate](/images/collate-logo.svg) Mayur Singal Software Engineer, Collate A walkthrough following a data engineer and steward at a bank as they build the context layer Collate AI runs on — quality investigation, AI-tagged classifications, governance review, and an agent that automates PII review against policy. Product Demo · Track 2 · Part 1 ### Consuming Context: From Question to Insight ![Collate](/images/collate-logo.svg) Pere Miquel Brull Software Engineer, Collate ![Collate](/images/collate-logo.svg) Nick Doyle CRO, Collate An analyst asks "Customer Lifetime Value by segment?" in natural language. Collate AI queries the warehouse and returns a chart grounded in governed metric definitions, all with visibility into provenance and logic. A business leader turns the analyst's chart into a weekly business process. Product Demo · Track 2 · Part 2 ### Consuming Context: Insight to Action ![Collate](/images/collate-logo.svg) Pere Miquel Brull Software Engineer, Collate ![Collate](/images/collate-logo.svg) Mayur Singal Software Engineer, Collate See how your open context layer can be used by different teams and systems. This demo shows how a data engineer can bring this context into Databricks, or into custom AI applications and services with the AI SDK. Product Demo · Track 2 · Part 3 ### Consuming Context: Governing Every AI Agent ![Collate](/images/collate-logo.svg) Sriharsha Chintalapani CTO & Co-Founder, Collate Bringing context into external systems via MCP, and registering every LLM, MCP server, and agent in the AI Governance asset registry. Bring your open context layer to govern your broader AI ecosystem. OpenMetadata milestones and award winners ## Closing: Community Update and Customer Awards Closing ### OpenMetadata Community Update and Customer Champion Awards ![Collate](/images/collate-logo.svg) Steve Wooledge Head of Marketing, Collate ![OpenMetadata](/images/omd-logo.svg) Shawn Gordon Developer Relations, OpenMetadata The closing session: OpenMetadata by the numbers (130+ connectors, 450+ contributors, 14K+ stars, 13,500+ community members) and the inaugural Customer Champion Awards recognizing OpenAI, Scout24, Ambry Genetics, Yelp, and Divisions Inc. ## Ready to put context to work? See Collate 2.0 in action on your own data. [Book a Collate demo](/book-demo)