Dataiku unveils orchestration-focused AI Success platform

Dataiku has launched what it calls the Platform for AI Success, a new version of its enterprise AI software aimed at companies moving beyond small-scale pilots. It is positioned as an orchestration layer that connects data platforms, enterprise systems, foundation models and third-party agent frameworks under a single governance approach.

The release also adds three products: Dataiku Agent Management, Dataiku Cobuild and Dataiku Reasoning Systems. Dataiku describes them as components that work alongside its existing platform and extend how organisations build, deploy and monitor AI agents and broader decision workflows.

Many large organisations now run AI across multiple cloud services, data environments and software vendors. As the number of tools grows across departments, approaches to monitoring, model selection and risk controls have become inconsistent. Orchestration has therefore become a priority, particularly for companies trying to standardise how systems are validated and monitored after deployment.

Dataiku says the platform connects to multiple systems without requiring customers to commit to a single vendor stack. The company presents this as a way to reduce vendor lock-in while maintaining central oversight of AI performance. The platform also includes tools to build, validate, deploy and monitor AI systems in a governed environment.

Three dimensions

Dataiku frames the Platform for AI Success around three themes: people, orchestration and governance. "People" focuses on how domain experts, analysts and engineers contribute. Orchestration coordinates data, models, agents and decision logic. Governance covers visibility, validation and performance measurement across both design and operational stages.

At the centre of the launch is Dataiku Agent Management, offered as a standalone product. It targets a gap that becomes more apparent as organisations deploy more AI agents across business functions: many monitoring tools can confirm an agent is running, but not whether it is delivering useful outcomes against business metrics.

Dataiku says Agent Management provides cross-platform visibility for agents regardless of where they are created or run. It evaluates agents against defined business KPIs and flags performance drift and cost issues. It can also trigger governance workflows based on risk thresholds and regulatory requirements.

Agent Management is available through an early access programme. Dataiku has not disclosed pricing or a timeline for general availability.

Reasoning systems

The second product, Dataiku Reasoning Systems, focuses on orchestrating multiple agents and decision components within a single operational environment. Dataiku describes it as a way to combine data, models, business rules and human-defined decision logic with governance and oversight.

The approach reflects an industry shift away from single-purpose agents operating in isolation. Many organisations now need AI systems that connect to data pipelines and integrate with existing operational rules. They also need clear accountability for automated decisions, particularly in regulated sectors and safety-critical workflows.

Dataiku says the first packaged offering is the Dataiku Reasoning System for Manufacturing Operations, available now. Planned Reasoning Systems for Supply Chain and Financial Risk are expected later in 2026.

Cobuild launch

The third product, Dataiku Cobuild, is scheduled for release in June 2026. It uses natural language prompts to generate what Dataiku describes as a complete AI project within its visual interface. The generated project can include pipelines, models, agents and applications, presented as governed workflows.

Dataiku positions Cobuild as an alternative to AI coding assistants that generate scripts. It says Cobuild creates a structured visual flow that users can review step by step before deployment. The company adds that its execution engine handles environment configuration, resource provisioning and deployment through a controlled process.

Chief Executive Officer Florian Douetteau said the platform was built around practical constraints organisations face when they try to scale AI beyond experiments.

"To achieve true AI success, enterprises face a critical conundrum", said Florian Douetteau, Co-Founder and CEO, Dataiku. "Without bringing everyone into the building process, AI initiatives won't be relevant or accepted; without orchestrating complex, modern technologies, AI will be too naive to have a meaningful impact; and without governing AI at every single step, it will never move beyond the proof-of-concept phase. We built our platform specifically to solve this exact roadblock."

Co-Founder and CTO Clément Stenac said the company sees orchestration and control as core requirements for AI systems that make operational decisions across a business.

"No amount of prompt engineering replaces structured orchestration," said Stenac. "Real enterprise decisions require data feeding models, models informing agents, and agents controlled by a necessary combination of explicit business rules and human oversight. That coordination layer is missing in most deployments, so the Platform for AI Success is designed to fill that void."

Dataiku says it will showcase the Platform for AI Success at the Gartner Data & Analytics Summit in Orlando, where it plans to demonstrate its approach to governed AI systems and measuring business outcomes from AI agents.