Fewer than one in five organisations keep a complete, current inventory of the AI systems they run, according to IBM’s “AI in Motion” research. Dataiku says that gap is exactly what its newly launched Agent Management product is built to close.
Key Takeaways
- Fewer than one in five organisations maintain a complete, current inventory of their AI systems, according to IBM research cited by Dataiku.
- Agent Management connects to AWS Bedrock, Databricks Agents, Google Vertex, Microsoft Copilot Studio and Azure Foundry, Salesforce Agentforce, Snowflake Cortex and Dataiku itself, plus OpenTelemetry for custom builds.
- The product was unveiled at Dataiku Succeed and became generally available in October 2026, priced per instance annually with monitoring metered per agent.
Dataiku unveiled Agent Management at Dataiku Succeed, its annual flagship conference, pitching it as a response to how quickly enterprises have adopted AI agents without any real way to track them. Most agent platforms, the company argues, can only see the agents built on their own infrastructure, leaving the rest of a company’s agent estate without clear ownership, purpose, risk rating, or measurable business outcome.
Why enterprise AI governance is lagging behind agent adoption
The scale of that blind spot is the point Dataiku keeps coming back to. Teams across a typical enterprise now build and deploy agents on whichever platform suits the task at hand, and no single one of those platforms has visibility into what the others are running. The result, according to Dataiku, is that most companies cannot say with any confidence how many agents they have in production, what those agents touch, or who is accountable for them.
“Ask a bank how many servers it runs, and you get an answer to the decimal. Ask how many AI agents it’s running, and you get a shrug or a guess,” said Florian Douetteau, co-founder and CEO of Dataiku. “Nobody set out to build it this way. Teams built agents faster than anyone could count them. Agent Management tells you what’s actually out there, and what it’s actually worth.”
How Agent Management finds and scores AI agents
Agent Management connects to the platforms enterprise teams already use to run agents, including AWS Bedrock, Databricks Agents, Google Vertex, Microsoft Copilot Studio and Azure Foundry, Salesforce Agentforce, Snowflake Cortex, and Dataiku’s own tools, with OpenTelemetry support covering custom-built environments. It scans all of them into a single inventory and automatically maps each agent’s structure, down to the specific tools and models it relies on.
For the agents rated highest risk, those handling customer interactions, sensitive data, or live transactions, the product keeps a standing record of certification status, named risks, and tests that rerun on a set schedule. The aim is that when a manager, auditor, or regulator asks for evidence, it already exists rather than needing to be pulled together after the fact.
A vendor-neutral approach to agent management
Dataiku is positioning Agent Management against the monitoring tools built into individual platforms, which it says tend to favour whichever vendor built them and see only their own slice of a company’s agent estate. Agent Management instead sits above the stack, aiming to answer questions no single platform can on its own: which agents are running unmonitored, where risk is concentrated across the business, and which agents are actually worth what they cost.
Availability and pricing
Dataiku says Agent Management became generally available in October 2026. Product access is priced per instance annually, with monitoring metered per agent on top.
Dataiku describes itself as the platform for AI success, built to connect enterprise AI stacks across multiple vendors under one layer of governance. Agent Management extends that pitch to the growing number of autonomous agents already sitting inside the businesses it serves.
