Feature
White Paper
Why Enterprise AI Fails at Scale
AI adoption starts at the top. But several studies examining executive use of AI show mixed results.
Ever since ChatGPT came out, organizations looking at artificial intelligence have been told to start with management. And according to the National Bureau of Economic Research (NBER), that’s true.
“AI use by executives is positively correlated with AI adoption at firms,” said Ivan Yotzov, a research economist at the Bank of England, one of the survey researchers. “Current AI adoption is around 40% at firms where the executives do not use AI at all. Meanwhile, it is above 70% at businesses where executives personally use these technologies themselves during the working week, and around 85% at businesses where executives use AI for more than 5 hours a week.”
Yay, right? Not exactly.
Table of Contents
- AI Reaches the C-Suite — But Not the Bottom Line
- The Next Phase of AI Adoption Is About to Scale
- What Are Executives Using AI for?
- AI Might Make Executives More Confident — And More Wrong
- AI Adoption Requires More Than Trust
AI Reaches the C-Suite — But Not the Bottom Line
The NBER report surveyed almost 6,000 executives from the United States, the United Kingdom, Germany and Australia, and found 66% of them use AI. But they use it for an average of only 90 minutes a week — and 80% said they saw no impact on either employment or productivity.
“This may be because these new technologies are still in a pilot/experimental phase within firms, they are still not widely used by all employees and there is also likely to be a learning component in integrating these new tools within business processes,” Yotzov said.
Executives were also more likely to use AI than younger workers, according to the 16th Annual Pulse of Talent report from Dayforce Inc. “While Gen Z brings digital fluency, executives are adopting AI faster because they are closer to the strategic imperative and see its impact on growth, productivity and competitiveness,” said Amy Cappellanti-Wolf, chief people officer at Dayforce.
But NBER found executives expected sizable impacts over the next three years, including 1.4% productivity boosts, increasing output by 0.8% and cutting employment by 0.7%. What’s going to change to make that happen?
Related Article: Executives Think They're Further Along in AI Than They Are
The Next Phase of AI Adoption Is About to Scale
“Intensity of use within businesses will change, as these new technologies become more integrated in existing businesses processes, adopted by more workers at scale and/or help create new ways of working,” Yotzov said. “Organization learning will also likely be part of the mechanism.”
Deloitte research backs this up. The company’s 2026 State of AI in the Enterprise report — an annual survey of 3,200 global business and IT leaders — found AI moving from pilots to enterprise scaling, said Beena Ammanath, executive director of the Global Deloitte AI Institute.
“While only 25% of respondents said their organization has moved 40% or more of their AI experiments into production to date, 54% expect to reach that level in the next three to six months,” she said.
For example, 34% of companies said they are starting to use AI to transform their businesses and 30% are redesigning key processes around AI, though the rest use AI only at a surface level. “Ironically," said Ammanath, "the same transformational potential that makes AI so compelling is also what makes it difficult to adopt."
In addition, 74% of companies said they plan to deploy agentic AI within two years. All of that could generate the effects the NBER respondents said they expected.
What Are Executives Using AI for?
What are executives doing with AI? The Dayforce report, which surveyed 6,954 employees, of whom 1,267 identified themselves as executives, gives some ideas.
- Writing & Editing: 63% executives, compared with 62% total respondents
- Admin Tasks, such as note taking and speech-to-text transcription: 65% executives, compared with 57% total respondents
- Research: 62% executives, compared with 54% total respondents
- Internal AI Agents, such as using a chat feature to ask HR questions or find paystubs: 57% executives, compared with 38% total respondents
- Image & Video Generation: 46% executives, compared with 30% total respondents
Another difference? Executives are three times more likely than workers to say their skills are becoming obsolete due to AI (45% vs 15%), according to Cappellanti-Wolf — though they didn’t specify which skills.
AI Might Make Executives More Confident — And More Wrong
AI-using executives might not actually be a good thing. A survey conducted by three business school professors found executives who used AI made worse decisions.
The researchers asked more than 300 managers and executives to make stock predictions after reviewing past trends. Half could then confer with their peers, while the other half could consult with ChatGPT. Participants could then revise their predictions.
Not only did executives who consulted ChatGPT make worse decisions than the other executives, they made worse decisions than they did on their own, the study found. ChatGPT made executives significantly more optimistic in their forecasts, while peer discussions tended to encourage caution.
“Like all experimental studies, we had to be specific and could not test all kinds of predictions like making a make or buy decision or about an M&A bid,” said Patrick Reinmoeller, professor of strategy and innovation at IMD in Switzerland. In more important decisions, this might be different, he acknowledged. “However, even if this were the case, this would not necessarily mean that such effort would make them immune to the effects that we found in our study."
It’s true that AI models are more advanced now and could make more accurate predictions (in fact, potentially influenced by reports of the study itself). And each change in the model requires retesting, as the original study conditions are now altered, Reinmoeller noted. But that may not help either.
“We found effects that are not based on the technology but on people’s cognition and behavior,” he said, attributing it to the “awe” people often feel about AI.
Related Article: ‘AI Psychosis’ Is Real, Experts Say — and It’s Getting Worse
AI Adoption Requires More Than Trust
The Dayforce survey highlighted similar cautions — 43% of execs said they'd be comfortable letting AI message their boss on their behalf without reviewing it, compared to 24% of managers and 9% of workers.
“This suggests a higher level of comfort with AI among executives, which may carry risks if not paired with appropriate oversight and guardrails," said Cappellanti-Wolf.
Webinar
Apr
14
The State of Enterprise Site Search: Moving Beyond "Good Enough"
Join CMSWire and SearchStax for a conversation about how enterprise IT and marketing leaders are moving beyond basic site search.
Register
Webinar
Apr
23
AI for Your DXP: Connect What You Have, Transform How You Work
Most AI strategies stop at the platform—but work happens elsewhere. Bring intelligence into Teams, email, tickets and CRM.
Register
Webinar
On demand
Content Leaders Collective: Navigating Content Decisions at Scale
Discover how content leaders are modernizing content operations, avoiding costly missteps and preparing for scale and AI.
Watch Now
Webinar
On demand
Do More with Less: Modernizing the Cloud Contact Center for 2026
Learn how to leverage cloud platforms without adding a single hire to personalize every customer interaction.
Watch Now
Webinar
On demand
Content Strategy Leaders Live: Scaling for Speed, Complexity and AI in High Tech
A candid roundtable on how high-tech leaders are rethinking content at scale.
Watch Now
Webinar
On demand
Cut the Noise: Deploying AI That Actually Moves the Needle
Learn how to turn AI experimentation into concrete revenue operations.
Watch Now
How should executives protect themselves?
“Considering the effects of LLM use on oneself and one’s team and organization is a must for leaders,” Reinmoeller said. He recommended that executives and employees be AI literate and understand how AI agents might interact. “Leaders of companies need to not only learn about AI but familiarize themselves through direct experience."
And they don’t have a lot of time, he warned. The first graduates who had AI at their disposal throughout their studies are entering the job market. “Companies need to be ready to make the best use of their enthusiasm for AI,” said Reinmoeller. “People and work are changing very fast. Organizations need to match that.”
But again, this may not help executives make better decisions.
“We hope that they check for this possible bias when interacting with LLMs and strengthen their defenses,” Reinmoeller said. “However, the astonishing achievements of ever better LLMs, the general hype around each update and the successes’ dominance of media time may at the same time ‘naturalize’ LLM usage and erode these defenses.”