Clients split on AI in consulting firms

Nearly every company is willing to let consultants use AI on their projects—just not on their own teams.

A new report found that 98% of businesses are open to consulting firms deploying artificial intelligence when delivering solutions. Yet 61% of those same companies report pushback against internal AI initiatives, sometimes to the point of outright bans in certain departments.

Where clients draw the line

The resistance isn’t evenly distributed. Finance, strategy, IT, marketing, and tax work top the list of functions where executives are most hesitant to adopt AI, with about 37% in each area expressing reluctance. Legal concerns, the cost of errors, and fears of over-reliance on external providers drive much of the skepticism in finance.

Cybersecurity and regulatory compliance face the least opposition, with only 26% of leaders resistant. The report, based on a survey of 150 senior executives who buy professional services, suggests that where mistakes are easier to spot or carry higher stakes, AI adoption slows.

Comfort levels with AI also vary widely. One in five companies will only accept extensive AI use if a human consultant oversees the outputs. Another 32% want AI limited to “discrete” tasks, avoiding anything critical. Meanwhile, 17% would actually prefer interacting with an AI agent over a person, citing its 24-hour availability and unlimited patience for questions.

Related: Grant Thornton CEO frustrated by inclusion business case

Consultants caught in the middle

The findings highlight a contradiction: clients want consultants to leverage AI for efficiency but remain wary of its risks when applied internally. Tony Maroulis, a senior consultant at Source Global Research, noted that while companies accept AI’s growing role in professional services, firms can’t rush toward fully automated solutions.

“AI is not infallible,” Maroulis said. “Consulting firms will need to maintain strict controls, especially as clients adjust to new ways of working.”

This tension isn’t new. A decade ago, cloud computing faced similar skepticism—companies adopted it for vendors but resisted migrating their own systems. The difference now is speed. AI’s rapid evolution means consultants must handle client expectations that shift faster than internal policies can adapt.

The report didn’t specify how many companies have formal AI policies, but the uneven adoption suggests many are still improvising. Some may be waiting for competitors to stumble before committing, while others are testing AI in low-risk areas to build confidence. Either way, the gap between external acceptance and internal resistance shows no signs of closing soon.

For now, consultants are left balancing two demands: delivering AI-powered efficiency while reassuring clients that humans remain firmly in control.