Just 6% of organizations express complete faith in AI agents to independently manage their primary operational functions, revealing a significant disparity between excitement for artificial intelligence and assurance in deploying it for the most vital procedures.
TL;DR
- Only 6% of organizations fully trust AI agents for primary operations, showing a gap between AI excitement and deployment confidence.
- Most businesses limit AI agent use to restricted, customary, or overseen duties due to concerns about financial, client, or employee impact.
- Despite readiness gaps in technology and data, AI implementation is rapid, with 86% anticipating increased spending on agentic AI.
- Security, privacy, and data quality are major hurdles, leading many to adopt "enterprise orchestration" for controlled AI integration.
This is merely one of the principal conclusions from a recent Harvard Business Review Analytic Services research report, commissioned by Workato, an automation platform vendor, and Amazon Web Services, which polled 603 executives and IT professionals globally in July 2025.
The analysis reveals that although agentic AI—which refers to systems capable of making choices and performing tasks with little oversight from humans—is becoming widespread, confidence is still largely limited to less critical responsibilities.
A significant portion of those surveyed, specifically 43%, indicated they only place confidence in AI agents for restricted or customary operational duties, while 39% limit their application to overseen scenarios or secondary functions. This trend highlights a reserved approach; the majority of businesses are open to trying new things but haven't yet reached a point where they're comfortable allowing AI agents to make independent choices impacting financial matters, clientele, or employees broadly.
Adoption outpaces readiness
Notwithstanding this warning, implementation is progressing rapidly. Nine percent of companies indicate they've completely rolled out agentic AI, and half report testing or investigating applications, with just 10% opting not to proceed following preliminary evaluation. Eighty-six percent of those surveyed anticipate a rise in spending on agentic AI within the upcoming two years, suggesting that trials will intensify even while doubts remain.
However, the research indicates that foundational abilities are falling behind. A mere 20% report their technological setup is completely prepared to back agentic AI for essential operations, 15% express the same sentiment regarding their information and frameworks, and only 12% believe that safeguards for risk and oversight are entirely established.
The report, employing a combined readiness metric that covers infrastructure, data, cybersecurity, and governance, categorizes 27% of entities as “leaders,”, 50% as “followers,”, and 24% as “laggards.”.
Advantages are tangible, yet they fell short of projections
Companies adopting agentic AI are already seeing measurable gains, but most have not yet achieved the full benefits they anticipate. Among organizations that have deployed or are piloting agentic AI and report any benefits, realized gains in these areas are lower than expected, though improved productivity, cost reduction, and customer experience still top the list. The data suggests that without robust foundations in data quality, governance, and architecture, organizations risk falling into “garbage in, garbage out” scenarios that erode trust in AI rather than build it.
Concerns regarding security and privacy stand out as the most significant obstacles to broader implementation. A substantial 31% of those surveyed identified cybersecurity and privacy issues as a primary hurdle, with apprehension about the quality of data output (23%), unprepared business operations (22%), and constraints in technological infrastructure (22%) following closely. Leaders are also contending with issues of governance, ambiguity in regulations, and possible effects on the workforce, which contributes to a preference for keeping AI systems distant from customer interactions or crucial operational tasks.
In reaction, numerous entities are adopting “enterprise orchestration,” which links together systems, information, and software into a controlled framework capable of securely fueling AI agents broadly. Eight percent indicate they've already deployed enterprise orchestration in readiness for agentic AI, and 74% are either developing it or intend to, with frontrunners significantly outpacing those falling behind. Over 80% consider offering access to applications and relevant details from data as highly or somewhat significant results of these orchestration initiatives.
The human element and the path forward
In addition to technological considerations, the document emphasizes individuals and the process of adapting to new systems as crucial elements determining if AI agents will transition from peripheral roles to central functions within an organization.
“The change management and reskilling that is going to be required across every company is something I feel has been underestimated,” said Kim Huffman, chief information officer at Workiva.
A substantial 44% of companies are placing a premium on educating or enhancing their workforce's skills in managing agentic AI, while 39% are concentrating on establishing robust guardrails and governance structures for responsible AI. Certain businesses are appointing AI liaisons and advocates across all departments to pinpoint potential applications and guide teams through initial trial runs.
Despite a mere 6% of businesses currently showing complete confidence in AI agents for essential operations, a substantial 72% of those surveyed believe the advantages of agentic AI surpass its potential drawbacks, and executives are particularly assured regarding this balance.
The research indicates that as organizations allocate resources to coordination, oversight, and employee preparedness, the confidence deficit within the company could diminish, potentially transitioning AI agents from trial assistants to reliable managers of the processes that shape organizational success.
For this story, Coins2Day reporters employed AI that generates content as a means of investigation. An editor confirmed the correctness of the details prior to its release..
