Today, Dataiku announced the launch of the Platform for AI Success — a major step forward in its mission to help enterprises move from experimental AI pilots to reliable, measurable business systems.

At FirstMark, we’ve partnered with Dataiku for years because the company has always been ahead of where enterprise AI was headed. Early on, that meant bringing data science and machine learning into a collaborative platform. Today, the challenge is much bigger: enterprises are trying to manage entire ecosystems of AI systems: models, agents, data pipelines, and decision logic, across increasingly complex environments.

The problem isn’t building AI anymore. It’s running AI as a reliable part of the business.

That’s exactly what this launch is about.

The Missing Layer in Enterprise AI

Across industries, AI is rapidly spreading across organizations. Foundation models, internal agents, third-party tools, and data platforms are being deployed everywhere, from customer support to manufacturing operations.

But the infrastructure to orchestrate and govern these systems at scale is still emerging.

Without a unifying layer, enterprises run into the same challenges again and again:

• AI systems are fragmented across tools and teams

• Business impact is difficult to measure

• Governance and oversight become inconsistent

• Costs and operational risk increase

In other words, organizations can launch AI initiatives, but they struggle to run them as durable systems that drive real outcomes.

Dataiku’s new Platform for AI Success is designed to solve this problem.

The platform acts as an independent orchestration layer for enterprise AI, connecting data platforms, models, agents, and enterprise applications in a single governed environment. It allows organizations to use the best technologies available (across clouds, models, and vendors) while maintaining centralized oversight and control.

This approach reflects a core belief we share with the Dataiku team: AI success will depend less on individual models and more on how well organizations coordinate AI systems across the business.

Three New Products Driving the Platform

The launch introduces three major new capabilities:

Dataiku Agent Management

As AI agents begin operating across enterprise systems, most monitoring tools focus on whether agents are running—not whether they are delivering meaningful results.

Dataiku Agent Management addresses that gap.

Designed as a standalone product, it provides cross-platform governance and observability for AI agents, regardless of where they were built or deployed. Organizations can monitor how agents are performing against business KPIs, detect cost or performance drift, and trigger governance workflows before problems affect operations.

In short, it shifts the conversation from “Is the agent online?” to “Is the agent actually delivering value?”

Dataiku Reasoning Systems

Many organizations are experimenting with standalone agents that automate individual tasks. But real business decisions rarely happen in isolation.

Dataiku Reasoning Systems orchestrate teams of agents working together within governed decision environments. These systems combine data, models, agents, and business rules into operational workflows that reflect how real organizations make decisions.

The first Reasoning System focused on manufacturing operations is available now, with additional systems for supply chain and financial risk expected later this year.

Dataiku Cobuild

Launching in June 2026, Cobuild is Dataiku’s new AI-assisted project generator.

Users can describe a business objective in natural language, and Cobuild generates a complete AI project inside Dataiku’s visual environment—including pipelines, models, agents, and applications.

Unlike “vibe coding” tools that generate opaque scripts, Cobuild produces a transparent visual workflow that teams can inspect, validate, and govern before deployment.

The goal is simple: make it dramatically easier for business experts and technical teams to build AI systems together.

From AI Activity to AI Performance

The deeper story behind this launch is a shift we’re seeing across the enterprise market.

For years, organizations have focused on AI activity: building models, experimenting with tools, launching pilots.

But the next phase of enterprise AI will be defined by AI performance: systems that are governed, measurable, and integrated into the way businesses actually operate.

As Dataiku co-founder and CEO Florian Douetteau put it:

“To achieve true AI success, enterprises face a critical conundrum: without bringing everyone into the building process, AI initiatives won’t be relevant or accepted; without orchestrating complex technologies, AI will be too naive to have a meaningful impact; and without governing AI at every step, it will never move beyond proof-of-concept.”

That orchestration layer (connecting people, systems, and governance) is what Dataiku has been building toward.

And it’s why we believe the company is uniquely positioned to help enterprises turn AI from experimentation into a true business performance engine.