Dataiku — Enterprise AI Orchestration and IDP Platform

Universal AI platform with document intelligence capabilities, preparing for a 2026 IPO with $350M ARR and a governance-by-design approach.

Overview

Founded in 2013 in Paris, Dataiku provides an enterprise AI platform that includes intelligent document processing through its Universal AI Platform. The company is preparing for a U.S. IPO in H1 2026 with Morgan Stanley and Citigroup as lead underwriters at a $3.7 billion valuation, down from a $4.6 billion Series E valuation in 2021, reflecting recalibrated market conditions. Dataiku reported $350M ARR in October 2025, up from $300M in January 2025.

In December 2025, IDC named Dataiku a Leader in its MarketScape for Worldwide Unified AI Governance Platforms 2026, its first major governance-specific analyst recognition. IDC Research Director David Schubmehl stated: "AI governance is no longer about policies and oversight alone — it has become an operational requirement embedded in how AI is built and deployed. Dataiku was recognized for its holistic approach to governance, compliance, and scalability, embedding these capabilities directly into the platform." IDC specifically cited Dataiku's enforced deployment controls that block noncompliant AI from reaching production, and its unification of DataOps, MLOps, and LLMOps governance in a single system.

By March 2026, Dataiku had moved decisively beyond its data science platform roots. The Platform for AI Success launch introduced three new products: Agent Management (cross-platform agent governance, Early Access), Reasoning Systems (industry-specific decision intelligence), and Cobuild (AI-assisted agent building, June 2026). The company simultaneously expanded its partner network into the Americas through a deal with Matrix, a top 10 global systems integrator with 17,000 employees.

Former Salesforce President Alexandre Dayon joined the Board of Directors in January 2026, adding enterprise sales depth ahead of the IPO.

How Dataiku processes documents

Dataiku's document processing runs through the Natif.ai IDP plugin, a modular pipeline that handles PDF, TIFF, and JPEG inputs using computer vision, deep learning, and natural language processing (NLP). The pipeline converts native and scanned content into structured data, with vision-language models (VLMs) extracting information from text, tables, and images in a single pass. Governance controls including audit trails, end-to-end traceability, and compliance checkpoints are embedded directly in the workflow rather than applied as post-processing overlays.

The Agent Hub extends this into multi-step agentic workflows: a collaborative workspace where AI agents can be built, shared, and scaled with ROI measurement attached. The AI Factory Accelerator, powered by NVIDIA, accelerates enterprise-scale deployments with native governance integration.

In February 2026, Dataiku moved its governance infrastructure into open source through the 575 Lab, its dedicated open-source office. Two toolkits are generally available: Agent Explainability Tools, which traces decision-making across multi-step agentic workflows and surfaces agent reasoning for data scientists, compliance teams, and end users; and Privacy-Preserving Proxies, which protects sensitive data end-to-end when enterprises use closed-source models, designed for local deployment. Licensing terms and GitHub repository URLs were not disclosed in available sources. Dataiku simultaneously joined the Linux Foundation and the newly formed Agentic AI Foundation. As CEO Florian Douetteau put it: "Enterprises need reusable building blocks that can become the standards for how agentic systems are controlled and inspected."

Platform for AI Success (March 2026)

The March 2026 platform launch is the clearest signal of Dataiku's strategic direction. CTO Clément Stenac framed the problem directly: "No amount of prompt engineering replaces structured orchestration. 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."

The three products address distinct gaps. Agent Management (Early Access as of March 9, 2026) provides cross-platform visibility and business-impact measurement for AI agents regardless of which system deployed them. It measures agents against defined business KPIs, flags performance drift and cost issues, and activates governance workflows based on risk thresholds and regulatory requirements. Reasoning Systems translates institutional knowledge into operational intelligence by coordinating data, models, agents, business rules, and human-defined decision logic in a single environment. Manufacturing Operations is available immediately; Supply Chain and Financial Risk are scheduled for later in 2026. Cobuild, launching June 2026, generates complete AI projects from plain-language business objectives in a visual, inspectable interface. Users validate step-by-step flows before rollout, which Dataiku contrasts with what it calls "vibe coding tools that produce opaque scripts."

Dataiku positions the full platform as a vendor-agnostic orchestration layer connecting data platforms, enterprise systems, foundation models, and third-party agent frameworks without dependency on any single cloud provider or supplier.

Use cases

Enterprise AI governance

Organizations use Dataiku's unified governance platform to close the gap where most enterprises cannot fully trace AI decisions end-to-end while AI is already embedded in daily operations. The platform embeds governance controls such as traceability, audit logs, and compliance checkpoints directly into AI development workflows. The 575 Lab's Agent Explainability Tools extend this to agentic pipelines, making multi-step agent reasoning inspectable by compliance teams without requiring custom instrumentation. IDC's assessment identifies fragmentation risk when different AI capabilities are governed by disparate point solutions, positioning Dataiku's unified approach as a differentiator.

Financial services and compliance

The Matrix partnership expansion to North America and Latin America targets financial institutions deploying AI-driven fraud detection, compliance, and enterprise risk solutions. Matrix VP of Data Services Gil Rozen noted: "For many financial institutions, modernizing risk and compliance systems has historically required lengthy, complex transformation programs. By combining Dataiku's AI platform with Matrix's advisory and delivery capabilities, we are seeking to strengthen fraud prevention, improve compliance, and scale enterprise AI adoption." The combined offering reduces deployment timelines from months to weeks. Dataiku Americas VP of Partnerships Taye Mohler added that the goal is helping "financial clients accelerate the deployment of AI-powered risk and compliance solutions while empowering teams across the organization to participate in building and scaling AI."

Retail AI transformation

Retailers use Dataiku's Retail Accelerator Pack for customer experience optimization and back-office automation. The pack includes seven ready-to-use use cases covering entity extraction and LLM-enhanced predictions. Head of AI Architecture Jed Dougherty notes the tension: "The riskiest place to use GenAI in retail is also the most valuable one: the customer experience." The accelerator is designed to compress deployment timelines for teams that cannot build from scratch.

Document intelligence workflows

Teams process document collections through the modular Natif.ai pipeline, converting native and scanned content to structured data with embedded governance controls for regulatory compliance and audit trails. The AWS Agentic AI and Healthcare Software Competency certifications extend this into healthcare-specific document workflows on AWS infrastructure, where compliance requirements are most stringent.

Technical specifications

Company information

Dataiku was founded in 2013 in Paris and maintains its US headquarters in New York City. Additional offices span Denver, Washington DC, Los Angeles, London, Munich, Frankfurt, Sydney, Singapore, Tokyo, and Dubai. The company raised a $200M Series F in 2022 at a $4.6 billion valuation (Series E, 2021) and is targeting an H1 2026 IPO at $3.7 billion with Morgan Stanley and Citigroup as lead underwriters. ARR reached $350M in October 2025, up from $300M in January 2025.

The 2026 Partner of the Year Awards reflect Dataiku's go-to-market structure: Snowflake, AWS, and Accenture took the three global slots, embedding Dataiku into enterprise data infrastructure and large-scale transformation programs before procurement conversations begin. No revenue or deal-volume metrics were disclosed for any winner.

Resources

• Website
• Natif IDP Plugin
• Document Intelligence Blog
• Knowledge Base
• 575 Lab Open Source