UK, US firms abandon AI plans amid skills shortage

General Assembly has found that most large professional services firms have abandoned at least one AI initiative because they lacked the necessary internal skills. The survey points to a gap between AI investment plans and workforce readiness across consulting, accounting and legal firms.

The research surveyed 258 leaders at director level and above in the United States and the United Kingdom, all at firms with at least 1,000 employees. It found that 61% of respondents had dropped at least one AI project in the past year, while 35% had abandoned multiple initiatives.

Skills shortages appeared most acute in some parts of the market. Among US firms, 65% had abandoned an AI initiative because of a lack of skills, rising to 73% among consulting firms.

The findings suggest many firms are pursuing AI mainly to improve efficiency in existing work rather than to rebuild their business around AI-led products. Seven in 10 respondents said their AI strategies focus on efficiency, while 30% are focused on building AI-first products.

Only 17% said they were moving towards a model where AI agents do most of the work. At the same time, 73% said they were either keeping headcount flat or hiring more people as they scale their use of AI.

That balance may not hold across the sector. Among consulting firms, 57% expect to invest more in AI-first products and solutions than in human augmentation in 2026.

Pricing pressure

AI is also affecting how firms talk to clients about fees. The survey found that 79% said AI is changing pricing conversations. Of those, 42% said clients are questioning their pricing model, while 37% said they are addressing the issue proactively.

Consulting firms reported the greatest pressure on pricing. This reflects a broader concern that AI may reduce the value clients attach to work traditionally billed on time, seniority or specialist knowledge.

The responses also showed a gap between confidence in AI adoption and confidence in AI-based services. While 88% of leaders said their workforce was very to extremely prepared to meet management's AI ambitions, and 94% believed they were ahead of their clients on AI adoption, only 26% said they felt "great" about selling their firm's AI solutions.

Most of the rest expressed more limited confidence. The survey found that 71% felt only "OK" about selling those services because AI does not always deliver as promised.

Partner gap

The survey suggests senior staff are a weak point in many firms' AI efforts. More than half of respondents, 55%, said partners were the least prepared group to use AI effectively.

The issue was especially sharp in legal firms, where 74% identified partners as the least prepared. By contrast, associates and analysts were most often seen as the best prepared, cited by 44% of respondents.

The biggest gap was not in technical model building or advanced engineering. Instead, 58% said the main shortfall was in change management and stakeholder communication, followed by prompting and using generative AI tools in daily work at 31%, AI governance, risk and compliance at 30%, and translating client problems into AI use cases at 28%.

A separate structural issue also emerged. Half of respondents said their partner compensation model makes it harder to invest in AI, rising to 57% at firms with more than 5,000 employees.

The results point to a tension between investment in tools and investment in people. One of the clearest signs is that firms are still relying on human judgment even as they automate more routine tasks.

"Professional services firms face a capability paradox," said Ash Khanna, Head of Professional Services at General Assembly.

Khanna described the challenge facing junior staff and firms' training plans.

"AI can accelerate output, but it can't replace the judgment that comes from doing the hard work. We're seeing junior talent generate recommendations faster than ever, but struggle to defend the reasoning behind them. Firms must invest in the durable human skills that make AI output credible and defensible. Upskilling can't be an afterthought; it has to be as central to a firm's AI strategy as the technology itself," Khanna said.

New roles

The survey also points to staffing changes as firms adapt to AI. Nearly two-thirds of respondents, 63%, said junior roles would evolve with different responsibilities and skill sets.

Over the next 12 months, respondents said they planned to recruit for AI-related roles including integrity layer leads, cited by 73%, workflow engineers at 47%, and agent orchestrators at 42%.

Khanna said the middle ranks of firms would have to change as AI alters the division of labour between junior employees and partners.

"Producing work is becoming a commodity, and the premium now is on accountability and judgment," Khanna said. "The firms that thrive will be the ones that help mid-level professionals–historically the QA layer between juniors and partners–reinvent themselves as workflow engineers: people who take the firm's unique, unscalable human expertise and industrialize it into a scalable digital asset," Khanna said.