Enterprises are struggling to maintain oversight as decentralized AI development outpaces centralized IT control, according to new research from Dataiku. At its Succeed 2026 conference in New York, the company addressed this friction by launching Agent Management, a standalone product designed to monitor and manage risk across AI agents regardless of their origin. This move comes as 84% of global CIOs report that employees are creating AI agents and applications faster than IT departments can govern them, signaling a critical need for orchestration layers in multi-vendor environments.
Perdue Farms and Air Canada Deploy Governed AI
Global organizations are moving beyond isolated AI experiments to integrate agentic workflows into core business operations. Perdue Farms is reportedly empowering business experts to act as AI builders by providing direct access to development tools. In the travel sector, Air Canada is utilizing agentic AI to scale the efficiency of its marketing operations. These use cases illustrate a shift toward democratized AI, where non-technical staff handle deployment, provided the underlying governance remains intact.
Other enterprise applications include Aviva, which consolidated 20,000 documents from over 3,000 audits into a single AI-powered chatbot to accelerate knowledge retrieval. Similarly, GE Aerospace is developing agents to capture and distribute decades of engineering judgment across its workforce. These implementations suggest that the current enterprise priority is converting specialized, often siloed, institutional knowledge into scalable, automated intelligence that remains subject to corporate oversight and data integrity standards.
Dataiku Expands Cobuild and Agent Management Capabilities
To address the "management challenge" identified in its Global AI Confessions Report, Dataiku is expanding its Cobuild natural-language agentic interface. The company is integrating its new Agent Management product directly into Cobuild, alongside the launch of Cobuild Insights and Dataiku Headless. These updates aim to provide a centralized orchestration layer that sits above existing cloud infrastructure and data platforms.
The company is also expanding its universal AI Catalog, which provides teams with vetted assets to build upon. This expansion targets a specific pain point identified in a study of 685 CIOs: 75% of IT leaders stated they cannot consistently measure whether their deployed agents are delivering actual business outcomes. By offering tools to discover and monitor agents across different environments, Dataiku is positioning its platform as a necessary layer for maintaining visibility in increasingly fragmented AI ecosystems.
Key Takeaways
- Dataiku's Global AI Confessions Report found that 84% of CIOs believe AI agent creation is outpacing IT governance capabilities.
- The company launched Agent Management, a standalone product for discovering and monitoring risk across AI agents.
- 75% of surveyed CIOs reported an inability to consistently measure the business outcomes delivered by their AI agents.
TechInsyte's Take
In our view, Dataiku’s launch of Agent Management is a direct response to the "shadow AI" problem currently destabilizing enterprise IT departments. As companies like Sonos and Perdue Farms move toward democratizing AI development, the risk of unmanaged, unmeasurable agents grows exponentially. Dataiku is betting that the market will shift from a focus on "building" AI to "orchestrating" it. By providing a layer that sits above multi-vendor environments, they are attempting to solve the visibility crisis where 75% of CIOs are currently flying blind regarding agent ROI. Success will depend on whether their governance can actually keep pace with the speed of decentralized creation.
Questions & Answers
How is the "governance gap" impacting IT leadership according to recent data?
The gap is significant; 84% of CIOs report that employees are deploying AI agents and applications faster than IT can implement necessary governance, creating a risk of unmanaged shadow AI within the enterprise.
What specific problem does the new Agent Management product aim to solve?
Agent Management is designed to help enterprises discover, monitor, and manage risk across AI agents, providing oversight regardless of the specific environment or platform where those agents were originally built.
How are enterprises like Aviva and GE Aerospace utilizing AI to manage institutional knowledge?
Aviva is using an AI-powered chatbot to consolidate 20,000 documents from 3,000-plus audits, while GE Aerospace is building agents to capture and distribute decades of engineering expertise across the organization.
Can enterprises currently measure the ROI of their AI agent deployments?
According to Dataiku's research involving 685 CIOs, 75% of IT leaders currently cannot consistently measure whether their deployed AI agents are delivering specific business outcomes.
Source: Dataiku
