Only 6% of companies fully trust AI agents to autonomously run core business processes, highlighting a clear gap between enthusiasm for artificial intelligence and confidence in unleashing it in the workflows that matter most.
This is just one of the key findings of the new Harvard Business Review Analytic Services research report, sponsored by automation platform provider Workato and Amazon Web Services, which surveyed 603 business and technology leaders worldwide in July 2025.
The report finds that while agent AI (systems capable of making decisions and actions with minimal human oversight) is rapidly gaining traction, trust remains severely limited to low-risk tasks.
For example, 43% of respondents say they only trust AI agents for limited or routine operational tasks, and 39% limit it to supervised use cases or non-core processes. This pattern emphasizes caution. While most companies are willing to experiment, they are not yet ready to let AI agents make unsupervised decisions that impact finances, customers, and employees at scale.
Adoption is outpacing readiness
Despite this caution, developments are proceeding rapidly. 9% of organizations say they have already fully implemented agent AI, half report piloting or considering use cases, and only 10% decide not to proceed after initial consideration. 86% of respondents expect investment in agent AI to increase over the next two years, suggesting further experimentation despite persistent skepticism.
However, research shows that basic functioning is lagging behind. Only 20% say their technology infrastructure is fully ready to support agent AI in core processes, 15% say the same about their data and systems, and just 12% feel that their risk and governance controls are fully in place.
Using a composite readiness index across infrastructure, data, cybersecurity, and governance, the report classifies 27% of organizations as “leaders,” 50% as “followers,” and 24% as “latecomers.”
Benefits are real, but less than expected
While companies implementing agent AI are already seeing measurable results, most still haven't achieved all the benefits they hoped for. Among organizations that have implemented or trialled agent AI and are reporting benefits, productivity gains, cost savings, and customer experience remain at the top of the list, yet realized benefits in these areas are lower than expected. The data suggests that without a strong foundation in data quality, governance, and architecture, organizations risk falling into a “garbage in, garbage out” scenario that undermines trust in AI rather than building it.
Security and privacy concerns are the biggest barriers to wider adoption. Thirty-one percent of respondents cited cybersecurity and privacy concerns as a top challenge, followed by concerns about the quality of data output (23%), unprepared business processes (22%), and technology infrastructure limitations (22%). Executives are also grappling with governance, regulatory uncertainty, and potential impacts on employees, reinforcing the trend to move AI agents away from customer-facing and mission-critical workflows.
In response, many organizations are turning to “enterprise orchestration,” which connects systems, data, and applications into a managed layer that can securely power AI agents at scale. Leaders far outpace laggards, with 8% saying they already have enterprise orchestration in place for agent AI, and 74% saying they are working on or planning to do so. More than 80% believe that connecting applications and providing context from data is a very or moderately important outcome of these orchestration efforts.
Human factors and the way forward
Beyond technology, this report focuses on people and change management as the deciding factor in whether AI agents move from the edge to the core of the enterprise.
“We feel that change management and reskilling is an underappreciated need for all companies,” said Kim Huffman, chief information officer at Workiva.
44% of organizations are prioritizing training or upskilling employees in agent-based AI monitoring, and 39% are focused on building responsible AI guardrails and governance frameworks. Some companies are appointing AI ambassadors and champions across departments to identify use cases and shepherd teams through early pilots.
While only 6% of companies currently express full trust in AI agents for core processes, 72% of respondents say the benefits of agent AI outweigh the risks, and leaders are most confident about that trade-off.
This research suggests that as companies invest in orchestration, governance, and workforce readiness, the trust gap at the heart of business may narrow, and AI agents may move from experimental helpers to trusted stewards of the workflows that define enterprise performance.
Regarding this story, luck Journalists used generative AI as an investigative tool. Editors verified the accuracy of the information before publication.
