---
title: "The Rise of the Agentic Data Catalog | On Demand | Atlan"
url: "https://atlan.com/context-in-practice-can-ai-document-your-data/"
description: "The traditional data catalog is dead. Prukalpa Sankar and data leaders from Mastercard, Sophos, and ASOS on how the agentic data catalog works in production. Watch the recording on demand."
---

> Atlan is hosting Context Conference, bringing together the leaders and builders at the frontier of giving AI the context it needs to understand their business. It runs online on October 28, 2026, from 11:00 AM to 2:00 PM ET. Atlan co-founder Prukalpa Sankar opens and closes the day. Leaders from AstraZeneca, BNY and Verizon share why they invest in context and what they get from it. Registrants get early access to The AI Context Gap, a new study from MIT Technology Review Insights. Register: https://atlan.com/context-conference/

**Traditional Data Cataloging Is Dead: The Rise of the Agentic Data Catalog.** An on-demand session from Atlan's Context in Practice series, recorded September 10, 2026, 58 min. Can AI really document your data estate? What does an agentic data catalog look like? How does governance evolve in an AI-first world? Prukalpa Sankar and data leaders from Mastercard, Sophos, and ASOS on making data AI-ready, and how an agentic data catalog generated **2M+ descriptions** and saved **210,000+ hours** across **200+ organizations**.

Watch the recording (free, email required): [https://atlan.com/context-in-practice-can-ai-document-your-data/](https://atlan.com/context-in-practice-can-ai-document-your-data/)

## Speakers

- **Prukalpa Sankar**, Co-founder, Atlan. Co-founded Atlan with a thesis that defined the third generation of data catalogs. Here she makes the case for what comes after: the agentic data catalog, built for AI as both the producer and the consumer of context.
- **Zhenni Hu**, Manager, AI and Data Governance, Mastercard. Led Mastercard's Context Agents rollout, enriching 30,000+ assets and saving 6,200 hours of steward time in two weeks, shifting her team from writing data descriptions from scratch to certifying what AI generates at scale.
- **Keith Guyett**, Solutions Architecture, Sophos. Enriched 40,000+ assets and built a repeatable scoring methodology, using a second LLM to evaluate output against five defined criteria, moving business users out of documentation and into a review role.
- **Rupal Sumaria**, Head of Data Management, ASOS. Enriched 13,000+ assets in two weeks with a team of two, then asked the next question: how does AI-generated context become the infrastructure layer for ASOS's own internal agents?
- **Nandini Tyagi**, Strategic Initiatives Lead, Atlan (moderator). Works with data leaders on how AI changes data cataloging and governance.

## What you'll take away

### Keynote / The Agentic Data Catalog (Prukalpa Sankar)

The catalog was built for one consumer: a person searching for data. Now the consumer is an agent, and everything about how context gets produced and consumed changes with it. What an agentic data catalog is, what it takes to build one, and why AI-ready data is the outcome that decides whether your AI program works.

- How the catalog category evolved, and what breaks when its primary consumer stops being a person and becomes an agent.
- Why every MCP demo looks the same, why no two behave the same in production, and the infrastructure it takes to close that gap.
- Where AI still doesn't work, how to organize around it, and the before-and-after mindset shifts from the teams that ran Context Accelerators.
- A first look at context from unstructured sources: what's already in beta, and where it goes next.

### Deep dive / The Mastercard Deep Dive (Zhenni Hu)

What it actually took for AI-generated documentation to meet the standard of the people who own the data. Not a proof of concept: a production rollout inside a heavily regulated financial services data estate, reviewed and signed off by the domain experts who had to live with it.

- **The Premise: when AI-generated context works.** Why it's viable now, and the line between enrichment your business will trust and enrichment that won't survive review.
- **The Build: Mastercard's approach.** A controlled scope first, stakeholders brought along early, and a team shifting from writing documentation to reviewing it.
- **The Outcome: what changes for the team.** A data estate AI agents can use, and a team doing judgment work instead of documentation work.

### Panel / Data Leaders Weigh In (Zhenni Hu, Keith Guyett, Rupal Sumaria)

AI-generated context looks convincing in a demo. Getting it to hold up with the subject matter experts who actually own the data is a different problem. Leaders from Mastercard, Sophos, and ASOS share what that took, and what it's changed about how their teams operate.

- **The Standard: when AI-generated context is good enough.** How four teams defined the quality bar, got their subject matter experts to agree on it, and what iteration actually looked like to reach it.
- **The Work: where the effort lived, and what's still unresolved.** Scope decisions, what fell short and why, and the mindset shift from treating this as a project to running it as a practice.
- **The Shift: what changes for governance.** When a team can speak directly to what AI can and can't do with your data, it starts showing up in different conversations, and earning a seat at a different table.

## What data leaders saw after the first review

- "We were stunned and perplexed by the quality of the content. How could the agents create such high-quality context from lineage, SQL, and dbt logic?" - Kenneth Jebjerg, Head of Data Engineering, Baader
- "The output shifted from solid generic descriptions to something that felt like it had been written by someone who understands our business." - Izabela Wilczynska, Data Governance Officer, PayU
- "I expected boilerplate, instead it inferred business context I never explicitly provided, correctly describing how an asset fit into our customer journey just from column names and lineage. That's when I realized this was knowledge synthesis, not just documentation." - Swatilekha Saha, Data Architect, DAT Freight & Analytics
- "Without this context we would have spent months/years manually updating the metadata required for these efforts and would not let us move at the speed we can now." - Bernie Daley, Director of Data Management, Nelnet Servicing, LLC
- "Before we started, 3% of our data estate had a description. It's now over 92%. I'd gone in expecting to correct most of what came out. I ended up mostly just checking it." - Greg Gaspar, Data Governance, The Bancorp
- "Before you spend another sprint tuning prompts, check whether your agent has any way of knowing what your data means. Ours didn't. By the end of the sprint we'd enriched 1,220 assets and the quality on the ones I could personally vet held up better than I expected. Two weeks moved that further than I thought possible." - Eshwar Vijay, AI Solutions Architect, Coupa
- "It did exactly what we needed. That sounds unremarkable until you've run an AI pilot. Most get you close to the goal, then eat a quarter in rework and stakeholder hand-holding. This one didn't." - Rodolfo Ramirez, Data Management, Workday
- "My assumption was that AI-generated documentation would be helpful, but still require significant human intervention. Instead, I found myself thinking less about documentation and more about a larger shift: context is becoming a strategic asset." - David Nishimura, Product & Program Leader, Itau Unibanco
- "Once we showed this to our BI team, you could see the attitude shift, from 'another tool, more work' to excited to get in and start looking around. The instant value just changed everything." - Sayali Avalakki, BI & Analytics Lead, Brightspeed

## Meet Context Agents

Your AI teammates that can write, maintain, and continuously evolve documentation. [Explore Context Agents](https://atlan.com/context-agents/).

| Agent | Role |
| --- | --- |
| Scout | Usage Intelligence |
| Scribe | Description Writer |
| Lexis | Glossary Builder |
| Doc | Readme Author |
| Nexus | Terms Linker |
| Sage | Metric Arbiter |
| Atlas | Domain Classifier |
| Vera | Quality Scorer |
| Orion | Ontologist |

## Watch the full session on demand

Context in Practice is a live series: real teams shipping AI-generated context inside regulated enterprises. This one ran September 10, 2026. The recording unlocks after a short email form on the page: [https://atlan.com/context-in-practice-can-ai-document-your-data/](https://atlan.com/context-in-practice-can-ai-document-your-data/)