Ataccama named a Visionary in the 2026 Gartner® Magic Quadrant™ for Data and Analytics Governance Platforms: Governing data for AI agents, and what comes next
That mandate lands on teams whose AI ambition has run ahead of the data underneath it. Many started with modernization, moving data to Snowflake or Databricks, yet their estate still spans on-premises and operational systems, and data that wasn’t fixed before it moved is stalling the AI projects built on top of it. Meanwhile, everyone in the business now asks AI for answers, and agents act on each other’s output, sometimes with no human in the loop. Untrusted data becomes untrusted AI, and bad automation at scale.
Today Ataccama was named a Visionary in the 2026 Gartner® Magic Quadrant™ for Data and Analytics Governance Platforms. In the previous edition, we were positioned as a Niche Player. So what did we change in our platform in the past year? A lot of course, but all with a clear strategy to support the mandate and the reality above: to be the independent data trust layer that enforces quality across the whole estate and integrates with the entire data stack to deliver trusted AI agents at the end of it.
Here are the principles behind that strategy and where we are taking the platform next.
Governance has to reach the whole estate
Many governance programs start from a reasonable plan: move the data to one platform and govern it there. In practice, enterprises run Snowflake, Databricks, legacy warehouses and operational systems side by side, and many keep a catalog they have invested in for years. A governance approach that assumes a single platform leaves most of that estate ungoverned, data issues remain unfixed at the source, and AI agents will reach into all of it.
Over the past year we opened Ataccama ONE to work alongside the estate you already have. You write a data quality rule once, and Ataccama enforces it where the data is – at rest or in motion: as pushdown SQL or a data metric function in Snowflake, as PySpark on Databricks, through a REST call, or at the edge inside on-premises systems, so regulated data never leaves your network. More than 200 connectors, plus any JDBC driver, extend that coverage well past the cloud platform.
Complex data stacks require interoperability
In December 2025, Snowflake Ventures invested in Ataccama. We read that investment as a signal that data cloud leaders see their customers’ data, and its trust problems, extending beyond their own platform coverage and support.
Following this investment we built multiple capabilities to help business and technical users work together and govern their Snowflake data as they would govern other data but while leveraging the benefits of elastic compute power and cortex AI with Ataccama’s MCP Server. Together with Snowflake and Deloitte we published a blueprint for trusted agentic AI.
We also built a converter for the OSI specification (now Apache Ossie) and contributed data quality metrics to it, so quality signals can travel with shared semantics. This ensures the governance work done on your data assets within Ataccama, can be reused across any system that uses the Apache Ossie format for their semantic layer.
Interoperability using smooth integrations and support for standards such as MCP and Apache Ossie is critical in order to match the fast pace an AI first world dictates and the existing reality of organizations on the other hand. This works hand in hand with the adaptability of the Ataccama platform. For example, you can keep the catalog you have and marry it with data quality scores you get from Ataccama, or consolidate both your metadata and quality work as a whole within Ataccama. You can co-exist or consolidate, on your timeline.
Agentic workflows should do stewardship work, with people approving it
Governance teams can’t keep pace with the volume of data that AI now puts to work. The answer isn’t bigger data teams but agentic support to take on part of the stewardship load. The ONE AI Agent works as your team’s digital data steward. It documents and classifies data in bulk, builds and applies data quality rules, and traces issues to their root cause before recommending a fix. A steward approves each change before it goes live.
Your team loads its own governance playbook into the agent as skills, so it follows your organization’s policies and conventions. Along with your custom skills, the agent draws on more than 50 tools built to support stewardship activities from cataloging, evaluating, monitoring and data referencing.
One large US utility company now governs 10x more datasets with 83% less manual effort, and its teams certify datasets in hours.
AI agents need a trust signal they can read
Semantics tells an agent what data means. Ataccama’s trust layer tells it whether to rely on that data.
The Data Trust Index is a single, machine-readable score that brings quality, ownership and context together for catalog items and data products. Agents read it through the Ataccama MCP Server along with more than 20 tools providing rich access to detailed quality and metadata context according to the user permissions. These insights are then available from inside Claude, Microsoft Copilot and Snowflake CoCo. When an agent can check trust before it acts, teams can move agents into production with evidence behind them.
This is exactly what customers like Lennar built with Ataccama:
“To us, it was almost immediate. What if we could tell you that the data that the agent’s telling you, the answer that it’s giving you, actually is correct because we validated it, because we have Ataccama’s data quality engine behind the scenes ensuring that this data is up to date, at par, and meeting the quality rules that we’ve set.” Isaac Gabay, VP of Data & AI at Lennar
Governance programs grow one domain at a time
Few governance programs begin with a full platform rollout. You can start with one area, such as data quality, catalog, reference data, MDM, observability or lineage, connect it to the rest of your stack, and grow into the platform as the program matures.
Mondelez started with data quality and expanded into a full catalog, business glossary and lineage management in Ataccama. Industrial Alliance started with MDM, then added reference data management so all of its entity resolution is governed in one place.
Licensing is built so you never get blocked. Volume limits are soft, there are no back-charges, and consumer users are free and unlimited.
What’s next: Getting you to trusted data faster
At FWRD 2026 on November 5, Chief Product Officer Jessie Smith shared where the platform is heading in 2027. Each direction extends the work described above.
- MDM with and for AI. Agents need one trusted version of your customers, suppliers and employees. We’re building agentic configuration in which Ataccama uses profiling to propose the data model and the matching and survivorship rules, and stewards review and approve them. The aim is MDM fast enough to deliver governed data products for AI without a multi-year program.
- Fast data quality. We’re building smart profiling that uses AI agents usage data to find your critical data elements and define your data quality foundation. Headless data quality will give MCP access to every operation, alongside a simpler UX, so data quality runs inside your own workflows, tools and agents. In addition, a new partnership with unstructured.io helps to augment Ataccama’s unstructured data governance abilities and cover more critical data for AI workloads.
- Agent trust. We’re working toward agent answers that each carry a trust score for the specific data slice in question in contrast to a quality score for the entire table that may not be granular enough to ensure trusted outputs. When the score is low, additional context explaining the possible reasons is exposed. The goal is AI answers you can trust, and fixed when they aren’t.
Customers can join early adoption for these capabilities through their customer success manager.
Recognized for vision and execution. Built to get you to trusted data faster.
Want to learn more about how we can help you achieve your data and AI governance goals?
FAQ
An agent acts on whatever data it can reach. Governed data gives it a quality and ownership signal to check first, so its answers rest on data the business trusts.
Ataccama is named a Visionary. In the previous edition, Ataccama was positioned as a Niche Player.
Ataccama works alongside the catalogs and cloud platforms you already run, and you can consolidate onto Ataccama ONE later if you choose.
A data trust layer sits across your data platforms and enforces quality wherever data lives. Ataccama adds a machine-readable trust score, the Data Trust Index, that AI agents can check before they rely on data.
Gartner, Magic Quadrant for Data and Analytics Governance Platforms, By Anurag Raj, Sarah Turkaly, Lulu Wang, Dalia Naguib, 12 November 2026
Gartner does not endorse any company, vendor, product or service depicted in its publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner publications consist of the opinions of Gartner’s business and technology insights organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this publication, including any warranties of merchantability or fitness for a particular purpose.
Gartner and Magic Quadrant are trademarks of Gartner, Inc., and/or its affiliates.
Jessie Smith
Jessie Smith is Chief Product Officer at Ataccama, bringing more than a decade of experience across enterprise data management, AI, and product development. She leads Ataccama’s product strategy, focused on helping enterprises improve data quality and build the trusted data foundation needed for AI.
