Most CIOs regret AI vendor, platform decisions: report

Dive Brief:

Nearly 3 in 4 CIOs regret a major AI vendor or platform decision made within the last 18 months, with 62% saying they’ve faced direct questions or challenges from their CEOs over vendor selection, according to a Dataiku report published Thursday. The Harris Poll conducted a survey of 600 CIOs globally on behalf of the technology vendor for the report.
A deadline is looming to prove the value of AI investments, as 71% of CIOs say AI budgets are likely to be cut or frozen if targets aren’t hit by mid-2026. Meanwhile, 85% of IT leaders say traceability and explainability gaps have delayed or stopped AI projects from reaching production.
“CIOs are moving from experimentation into accountability faster than most organizations expected,” Florian Douetteau, co-founder and CEO of Dataiku, said in a press release.

Dive Insight:

As CIOs face critical questions over AI vendor selection, they’re also dealing with how to manage employee use of AI tools.

More than half of the CIOs polled for the Dataiku report said they found employees using unsanctioned shadow AI tools, apps and platforms to complete work tasks. Unfettered AI access will create significant technical debt within an organization, according to 89% of CIOs.

Six out of every 10 corporate employees will use shadow AI tools if they’ll help them meet work-related deadlines, with 86% of workers using AI tools at least once a week to accomplish tasks, according to a report from security firm BlackFog published in January.

The Dataiku report revealed that employees are also creating AI agents and apps faster than IT is able to govern them, adding to agent sprawl, according to 82% of CIOs. As enterprises add AI agents to business-critical workflows, the governance gap is growing, the report found.

Agent sprawl is a pain point for IT leaders across industries as retail companies, food and beverage providers and more deploy the technology in a push to gain efficiencies and revenue. Yet research firm IDC predicted that the number of active AI agents will exceed 1 billion globally by 2029.

While 87% of CIOs said AI agents are embedded in critical systems, only 25% said they have full visibility into all agents in production, the Dataiku report found.

“The pressure is real,” Douetteau said. CIOs will need to build AI systems they can “explain, govern and stand behind before accountability is imposed rather than chosen.”