Opening Keynote
- Prukalpa SankarCo-founder

Virtual Summit · Dec 3 · 11 AM ET
Speakers from


Everything your catalog knows, including definitions, lineage, policies, and the business knowledge around them, can now live in an open, Iceberg-native store that any engine or agent can query. Abhishek Mishra walks through the decisions behind it and the architecture patterns emerging across the customers who've turned it on: Why it's designed to stay open and portable, how it keeps a record of exactly what context an agent saw on any given day, and how to get started.
Community speakers announced soon
Get a close look at a Context Lakehouse running at enterprise scale, and the blueprint behind it: The design choices the team made, how it fits alongside the rest of their stack, and how it puts governed context within reach of the AI their business is building.

People have stopped typing keywords to find data. They ask in plain language, in chat tools like Slack, and more and more from inside the AI tools they already use. Nandini Tyagi has built alongside more than 100 of the world's leading data and AI teams, and walks through what that shift means for adoption: Where people now go for answers, and how to meet them there.

Users will try a new way to ask about their data once. Whether they come back depends on the answer. With Conversational Search, anyone can ask from Slack, Teams, or their browser and get an answer grounded in governed context. Panelists who've rolled it out trade notes on how to keep a rollout from stalling, and what kept their people asking long after launch day.

MCP can carry your catalog's context into almost any AI tool, but getting it connected is the easy part. The value comes from what you do next: Trace a broken dashboard to its root cause, see what a SQL change will break before you merge, ground AI-written code in your real schema, score assets for AI readiness, or grow your business glossary. Steven Hloros opens up the MCP Cookbook, more than 30 tested recipes gathered from the ways customers put MCP to work, and explains how to contribute your own as it keeps growing.


Many teams already have MCP connecting their context layer to an AI tool or two, but the protocol reaches much further: AI assistants and internal chatbots, engineers' code editors, enterprise search, workflow automation, data pipelines, even agents that keep the catalog documented. Panelists building with it today share the use cases they've put to work, how they decided where MCP fits in their business, and what held up as they built it out.

Coverage tells you how much of your catalog is documented. It can't tell you whether that context actually worked. Every question people and agents ask leaves a trace, and Steven Hloros walks through how teams read those traces to find where answers went wrong, route each gap to the right fix, and make the next answer better than the last.

Getting an analytics agent to answer is quick. Getting it to answer with your business's own numbers, the ones the business already agreed on, is the hard part. Atlan's forward-deployed engineers do that work alongside customers, and Zach Michel shares the playbook they've built from it, so your team can start its next agent from what's already proven instead of from scratch.


Leadership wanted fast, specific answers across five businesses, and agents built straight on the tables kept guessing at the metrics. The definitions the business had already agreed on were sitting in its dashboards. The team from AD Ports Group shows how they turned those dashboards into the context their agents query, then tested every agent against the questions people really ask until the answers held.

BusPatrol's data team had become the place every question about its data ran through. Alongside its work to spread that knowledge, the team built a talk-to-data agent, starting as a side project and growing it one group of users at a time. Each group's questions showed which features and context to add next, and made one thing clear: A path for context is theirs to own, whatever tools win out.
Community speakers announced soon

Jessie BueltemanSr. Manager, Enterprise Data GovernanceElastic
Sushma RaoSVP, Data ManagementM&T Bank
Ravi VaishnavSenior Specialist, AI & MLAD Ports Group
Sonia SanchezIS Dept. Project LeaderAD Ports Group
Jason ScarlettExecutive Director, Enterprise Data EngineeringAlberta Health Services
Prukalpa SankarCo-founderAtlan
Gary DeffendallManager, Data GovernanceSouthState Bank
Jeremy LowellPrincipal ArchitectDAT Freight & Analytics
Simon KingVP, Data OperationsLoopback Health
Joseph NowickiSVP, DataBusPatrol
Abhishek MishraSenior Product ManagerAtlan
Nandini TyagiStrategic Initiatives LeadAtlan
Steven HlorosSr. Customer Education ManagerAtlan
Zach MichelVP, Customer Onboarding & SupportAtlan
See where each part of the Agentic Data Catalog is headed, how your peers put it to work, and what you need to start on your own.
Report on what's governed and what's ready for AI, straight from the engine you already use.
More than 30 tested recipes, gathered from the ways customers put MCP to work.
Read the traces every question leaves, find where answers went wrong, and make the next one better.
The playbook Atlan's forward-deployed engineers built alongside customers, so your next agent starts from what's already proven.