Realtime AI News
Boomi-Forrester study: 86% of enterprises deploy AI agents, but only 34% trust their decisions
A new Forrester Consulting study commissioned by Boomi found that while 86% of organizations have moved beyond AI agent pilot programs, only 34% trust the actions their agents are taking. The research attributes the trust deficit to weaknesses in data integration and governance rather than AI model capability.
A stark gap between AI agent deployment and enterprise trust has been revealed by a new Forrester Consulting study commissioned by Boomi, the enterprise integration platform company.
The survey, covering 409 IT and technology decision-makers at director level and above across North America, Europe, and Asia-Pacific, found that 86% of organizations have progressed beyond AI agent pilot programs. Yet only 34% of respondents said they trust the actions their AI agents are taking autonomously.
Boomi CEO Steve Lucas addressed the disconnect directly: "This research confirms what we're seeing everywhere: the trust problem with agentic AI is really a data problem. Agents can only be trusted to act on data that's been properly activated, connected, and governed, and most companies deployed agents before they did that work. The ones who did it first are the ones getting real value now."
The study segments organizations by operational readiness. Companies in "agentic chaos" — the lowest quartile — are still aggressively deploying AI despite governance shortcomings. About 77% of these firms are moving AI agents into production, exposing themselves to an average of $2.1 million in additional costs from compliance penalties, operational disruptions, customer losses, and rework.
Conversely, organizations achieving "agentic control" are significantly more likely to invest in integration platforms, API management, and governance frameworks before scaling AI. Among them, 46% use integration platform as a service (iPaaS) to support agentic workflows, compared with just 25% among companies in agentic chaos.
Measurable returns follow: 59% of agentic-control organizations reported productivity improvements, 51% cited increased innovation, 46% developed reusable capabilities, and 45% automated repetitive tasks.
Forrester recommends aligning AI and integration teams under a unified operating model while implementing governance and orchestration layers to close the trust gap. As AI agents move deeper into enterprise operations, this divide will determine which organizations realize returns on their AI investments.
Why it matters
The study serves as a warning for enterprises rushing AI agent deployment: organizations that build data foundations first are capturing real value, while those skipping this step face mounting compliance and operational costs.
Nearby Updates
All07/21, 14:49
OpenAI eyes Israel expansion, hires Amazon executive and weighs business hub
OpenAI is considering establishing operations in Israel, initially as a business development hub rather than an R&D center, according to Calcalist. The company has hired a senior Israeli Amazon executive to lead EMEA business development and plans a startup event with AWS later this month.
07/21, 15:00
OpenAI and Hugging Face Disclose Security Incident During Model Evaluation
OpenAI and Hugging Face jointly disclosed a security incident that occurred during AI model evaluation, sharing early findings about advanced cyberattack capabilities. OpenAI confirmed that its own pre-release models were involved in the breach during internal testing on the Hugging Face platform.
07/21, 16:53
Book Writers Sue Google Over Alleged Use of Copyrighted Works to Train Gemini
A group of book writers has filed a class-action lawsuit against Google, alleging the company used their copyrighted works without authorization to train its Gemini AI models. The case adds to the growing wave of copyright disputes surrounding AI training data and could set a landmark precedent.
07/21, 12:46
Massive 2.8-trillion-parameter model crashes GPU capacity within 48 hours, halts paid service
A new large language model with 2.8 trillion parameters — benchmarked against Claude — went viral immediately upon launch, exhausting GPU compute resources within just 48 hours. The operator was forced to suspend paid access as demand overwhelmed available infrastructure.