Artificial intelligence data analytics startup Dataiku Inc. today announced the launch of AI Agents with Dataiku, a new set of capabilities designed to create and control AI agents at scale, delivering AI applications powered by analytics, predictive models and agents.

The capabilities seek to assist companies that, as they rush to deploy agents, may create clumsy architectures where agents operate without proper IT control, vary in quality and relevance and are sprawled ungoverned across teams. Dataiku takes on the challenge by augmenting its Universal AI Platform with agents as true enterprise systems that are grounded in trusted data, embedded in operational workflows, connected to all AI inputs and governed with the same rigor as any business-critical asset.

AI Agents with Dataiku introduces two development tracks: a visual agent for nontechnical users and a code agent for developers. Both approaches are available within the same platform and allow cross-functional teams to collaborate without sacrificing control or flexibility. Each agent is governed using built-in validation workflows to ensure consistency, quality and compliance.

To support the secure orchestration of AI agents, Dataiku has implemented the “LLM Mesh,” architecture that manages access to models from providers like OpenAI, Anthropic, Mistral and open-source options such as Llama. The platform supports both self-hosted and cloud-based deployments across services, including AWS Bedrock, Microsoft Azure and Google Gemini, to allow companies to integrate AI agents into their preferred infrastructure while maintaining control over data residency and access.

Security and governance also get a look in with Dataiku Safe Guard, a feature that lets information technology teams apply flexible guardrails across agent interactions. Combined with Dataiku’s Agent Connect feature, organizations can centralize agent usage and dispatch requests to single or multiple agents as needed.

Central to the new offering is performance and observability, with tools such as Trace Explorer providing transparency into agent decision-making and debug workflows, while another feature called Quality Guard uses LLM-as-a-judge mechanisms to evaluate outputs.

Dataiku has also added a generative AI Registry for strategic oversight that allows business leaders to review agent use cases, assess risk and determine business value before agents are moved into production. The centralized registry supports broader efforts to align agent deployment with organizational priorities.

“As companies mature from exploring to operationalizing agents, they’ll quickly realize the need to centralize their creation for proper governance instead of letting them sprawl, continuously optimize their performance instead of letting them stale and fully orchestrate them through existing IT assets instead of raising technical debt,” said co-founder and Chief Executive Florian Douetteau.

Image: SiliconANGLE/Reve