AI agents may be making more headway in tech headlines than in actual offices.
That’s according to the latest adoption report from accounting firm and consultancy EY. It found that while 34% of the 500 senior business leaders they surveyed in April said they’d begun to try out agentic AI, only 14% had fully implemented generative AI systems that can act beyond the realm of a chatbot.
The biggest barriers holding back adoption were trust and safety. When it comes to handing over the reins of back-office functions to autonomous systems, respondents said they were worried about the cybersecurity implications (35%), data privacy headaches (30%), insufficient regulation (21%), and lack of company policies (21%). Nearly two-thirds (64%) said they think fears of AI replacing humans will prove an obstacle, though only 24% reported employee resistance as a barrier.
What agents are doing: Many of the business functions agents are currently tasked with are more marginal than what EY’s report authors think they can handle.
Agents are doing things like “assisting/ or managing processes,” “assisting tasks,” and “managing workflows,” respondents reported. But Dan Diasio, EY global consulting AI leader, said they should be doing more complex work.
“By their nature, agents are intended to autonomously complete tasks and even work with other agents, not merely assist or manage,” Diasio wrote in the report. “These activities are all useful, but they only scratch the surface of what’s possible. We’re talking about a step change in how work is done, not going a couple of steps beyond robotic process automation (RPA)-level tasks.”
Keep up with the innovative tech transforming business
Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.
Nearly three-quarters (73%) of respondents agreed, saying they can foresee agents managing entire business units. But nine in 10 said that built-in human intervention will always be necessary for agents in the wild.
“These statistics seem to contradict each other,” Diasio wrote. “This speaks to how we are both impressed with and intimidated by what we’re seeing from AI—how the technology forces us to reassess our own futures in the workplace and in society—as well as how many challenges on adoption persist, including just our own mindsets.”
We have AI at home: EY’s report also found that 64% of companies are now focused on building out in-house AI capabilities rather than acquiring existing firms, an increase from 56% about a year ago. The consultancy found that 32% of AI investments were devoted to buying “solutions off the shelf”—not customized for the company—a drop of 6% from six months earlier.
Investment in AI seems to be continuing to tick upward; a little more than a fifth of respondents said their companies have now invested at least $10 million in AI, and 35% planned to hit this milestone next year.
EY Americas AI and Data Leader Traci Gusher wrote that the shift toward in-house may reflect how many enterprise platforms are already building in AI features, making purchases of off-the-shelf solutions redundant.
“Buying a separate solution in the market off the shelf may seem unnecessary,” Gusher wrote, “and building a custom tool, to differentiate from the sea of sameness, may in turn feel imperative.”