AI Pressure Forces Professional Services to Rethink Value

([1])Accenture’s latest earnings sent shockwaves through the consulting sector this week. The firm posted decent fiscal Q3 results – revenue up 6% to $18.7 billion and solid margins – yet it slashed its growth forecast to 3–4% (from 3–5%), citing slowing new bookings and client caution. Investors punished the stock, which plunged about 18% in a single day, its worst one-day rout in years ([2]). The message was clear: even the best-run consultancy isn’t immune to market fears about AI-fueled disruption.

([3])At the heart of the concern is how generative AI threatens consulting’s traditional work (and revenues). A large share of consulting income comes from designing systems, scouring data, and other labor-intensive tasks – the kind of work typically leveraged through armies of junior staff. Now, many of those tasks can be accelerated or even automated by AI tools in a fraction of the time ([4]). This means clients may still hire advisors, but they won’t pay for huge teams to grind through coding, research, documentation, or basic analysis when an AI co-pilot can do the heavy lifting almost instantaneously ([5]).

Consulting firms are already adapting in visible – and painful – ways. Accenture’s CEO Julie Sweet insists that “AI will be a tailwind for us and our industry” as it scales ([6]), but the immediate reality is more complicated. One analyst noted that Accenture’s massive AI investments have "yet to translate into meaningful revenue acceleration" ([7]). In the meantime, firms are cutting costs and rethinking staffing: McKinsey reportedly shed 5,000 employees and Accenture 11,000 in the past year amid an industry shift to automation ([8]). Even the Big Four are hiring fewer entry-level consultants and more AI engineers as they acknowledge that software can replace junior grunt work ([9]).

([10])The uncomfortable irony is that the very AI solutions consultants are selling to clients could shrink the need for consultants themselves. As one analysis pointed out, the AI tools transforming what clients want from firms are the same tools “compressing the labor-per-engagement that underpins [the firms’] revenue model” ([11]). In other words, if projects require fewer billable hours thanks to automation, the legacy pyramid model – where partner profits depend on leveraging a large base of junior billers – begins to wobble. Unless consultancies can create new value streams (and pricing models) around AI, they risk tech-driven margin compression even as they tout efficiency gains for clients.

The response from leading professional service firms has been to go all-in on AI – both in external branding and internal operations – with mixed results. In one ambitious example, KPMG’s Hong Kong arm just released an Audit Quality Report highlighting a sweeping “Digitalisation of Audit” program. The firm has expanded a centralized Audit Platform and rolled out tools like “KPMG Clara AI Chat,” which lets auditors query documents and get AI-driven coaching from a secure knowledge base ([1]). KPMG also introduced an AI-powered Vouching Platform to automate routine audit verification work, boosting productivity by up to 40% in some cases ([2]) ([3]). The effort, part of a massive investment in AI and data training for thousands of staff, aims to modernize audits while bolstering trust in Hong Kong’s capital markets.

While KPMG races forward on transformation, it has also stumbled. In a cautionary tale for the industry, KPMG had to pull down a high-profile report on AI adoption after discovering that large portions were built on “hallucinated” case studies and statistics ([4]). Multiple organizations cited in the report (from UBS to the UK’s National Health Service) publicly rejected its claims as false ([5]). The firm admitted the report did not meet its quality standards and vowed to tighten controls, reminding employees that human oversight must validate AI-generated content ([6]). This episode – coming just weeks after a similar AI-induced report retraction at EY ([7]) – shows that even the biggest consultancies need to be extremely careful when using generative AI in client-facing analysis.

Other top firms are making high-profile AI bets while managing expectations. Deloitte, for example, announced this week the opening of a new "AI Studio" at its London office in collaboration with Google Cloud ([8]). Billed as a co-innovation hub, the facility will bring together Deloitte experts and clients to develop “agentic” AI solutions – software agents that can autonomously tackle complex, multi-step processes ([9]). The goal is to help enterprises move from piecemeal AI experiments to production-scale deployments that deliver tangible business outcomes ([10]). Such moves show the Big Four’s determination to be seen as leaders in AI. Yet, behind the scenes they also acknowledge that trust, governance and upskilling are critical. Deloitte is training 1,000 UK staff on Google’s latest AI platform Gemini and emphasizing ethics and oversight in all projects ([11]) ([12]). The big firms want to harness AI to improve margins and results, but they can’t afford another misstep that might erode client confidence.

The $900 billion global legal industry is equally caught in AI’s disruptive vise. Corporate clients are adopting generative AI to do work in-house, threatening to shrink a lucrative revenue stream for law firms. In fact, over 60% of corporate legal departments say they plan to rely less on outside counsel because of AI-driven efficiency gains ([1]). Many general counsels now explicitly expect their law firms to use AI to work faster and cheaper – yet paradoxically, less than one-third of clients even know if their outside lawyers are using AI at all ([2]). This misalignment is creating tension: clients want the cost savings of AI, but they also worry about quality and risk, leaving law firms walking a tightrope on when and how to inject AI into their matters.

On the supply side, law firms are scrambling to build their own AI capabilities to stay indispensable in the face of these shifts. Specialized startups are offering to help “bottle” a firm’s unique expertise into custom AI models. For example, Harvey – an OpenAI-backed legal AI company – is now working with top firms on training open-source language models with those firms’ internal workflows and client-specific playbooks ([3]). The idea is to capture how experienced lawyers handle complex, repetitive matters for key clients, and then automate large portions of those processes with precision. By encoding institutional knowledge into AI, firms hope to deliver faster service for routine work without eroding quality.

Big Law heavyweights are taking matters into their own hands – quite literally investing to develop proprietary AI. Leading firm Kirkland & Ellis made waves with a $500 million partnership with Palantir to develop an AI platform that encapsulates its coveted legal methods (or “secret sauce”) for certain high-stakes workflows ([4]). Data giant Thomson Reuters is similarly training open models on its vast trove of legal research and know-how ([5]), aiming to augment its Westlaw services with homegrown generative AI. These efforts reflect a broader strategy: if off-the-shelf AI might commoditize routine legal tasks, then capturing differentiating expertise in systems could help firms maintain an edge.

Meanwhile, entirely new breeds of tech-enabled legal service providers are emerging. ([6])New York-based startup law firm Crosby, for instance, has from day one abandoned the billable hour model. Its 30-lawyer team works alongside AI agents to review contracts in record time, and clients are charged a fixed fee per page instead of hourly rates ([7]). Crosby’s pricing can be as low as $10–$50 per contract page – a stark contrast to the thousands that traditional firms bill for complex contract reviews ([8]). With $60 million in venture backing and around 100 companies already on its roster, Crosby’s “AI-first” approach shows how professional services can be productized and scaled. Incumbent firms may loathe the idea of scrapping the hourly model, but competitors born with AI in their DNA are proving that new approaches can entice clients.

All these fast-moving developments beg a fundamental question: in a world where machines can handle so much of the grunt work, what is the role of human experts? Professional services have long sold human judgment as their core offering. That won’t change – but the nature of that human judgment is being redefined. Routine number-crunching, basic research, first-draft writing and data analysis are increasingly the domain of AI. The enduring value of a seasoned consultant, lawyer, or accountant will lie in the uniquely human elements: understanding context and nuance, exercising ethical and strategic judgment, and stitching together disparate pieces of insight into a coherent plan of action.

Crucially, human expertise is also the fail-safe that keeps AI on track. Clients and regulators will not tolerate AI errors that lead to bad decisions or compliance risks, so firms must ensure rigorous oversight of any AI outputs. The recent KPMG report fiasco hammered home this point – even a powerful model can invent convincing lies if left unchecked ([1]). As a result, the demand for professionals who can effectively supervise AI, validate its results, and articulate insights to clients is set to soar. Far from rendering experts obsolete, AI makes their high-level guidance more important than ever.

([2])As Hani Ashkar, a senior partner at PwC, noted, expert human judgment will “continue to shape the decisions that matter” – no matter how powerful technology becomes. In the end, AI may handle the heavy lifting, but human advisors will define the problems, ask the right questions, and ensure that technology’s outputs are turned into real-world strategic value. For forward-looking professional service leaders, the task now is to pivot fast: invest in AI for efficiency, yes, but double down on those human elements that machines cannot replace.