Consultancy firms have committed billions to frontier AI tools that have fundamentally changed tasks traditionally done by junior staff, yet the industry still cannot quantify what that spending actually returns.

The Big Four alone — Deloitte, KPMG, EY and PwC — have collectively committed several million pounds towards AI investment, according to industry figures, and the pace is accelerating across professional services.

At the centre of the shift is Model ML, a platform co-headquartered in London and New York that automates the manual backbone of consulting work: supporting research, building presentations and Excel models, and reviewing client-facing materials.

These are precisely the tasks that graduates entering the sector have performed for decades. A growing number of consultants now lean on the platform instead, upending a staffing model that has defined the industry since its modern formation.

What does the uptake look like in numbers?

A recent case study with BDO, one of Model ML's newest partners, quantifies the scale of adoption. Over the first eight weeks, the accounting firm's deal teams:

• asked more than 10,000 questions through the platform
• analysed over 25,000 documents
• generated more than 14,000 AI review comments

That volume, compressed into roughly two months at a single client, illustrates how quickly AI-assisted review and analysis has moved from pilot to daily practice in deal work.

Raghav Sehgal, a former Bain executive and now Model ML's general manager, said the technology does not eliminate junior roles but redefines them. Because consulting carries massive reputational risk once a client logo is attached, he explained, the job shifts — particularly for juniors — toward rigorous double-checking, insight creation, and managing the AI's output.

Where is the return on investment?

The biggest open question in consulting, as in other industries, remains the real ROI of these expensive tools. Sehgal breaks the equation into two parts.

"ROI ultimately is a function of two inputs," he said. "One is the tool itself … Does it fundamentally deliver and do what it's supposed to do … and two, it's change management: how do you reorganise yourself as a company around these tools?"

Milan Malik, former McKinsey consultant and head of consulting at Model ML, framed the same point more bluntly. "ROI is a function of capability and adoption," he said. "If the platform performs at the frontier, the only thing left to solve is adoption, and that's a change management question."

In other words: the technology works, but firms capture value only if they restructure how people work around it.

How are firms driving adoption?

Firms are already experimenting with incentives, Malik said. He pointed to two examples among Model ML's clients:

• One firm has placed its AI model on an internal learning platform, rewarding staff with learning credits if they upskill and share wins with colleagues
• Another publishes a weekly leaderboard of its top users across every region, which turns them "into AI champions and creates a competitive spirit that drives adoption and, ultimately, ROI"

Sehgal, a former consultant turned AI specialist, described "an exciting time" for the sector, because "the way you work with your clients internally is changing."

From finance to the Big Four

Model ML was founded by brothers Chaz Englander and Arnie Englander. The company started in the financial sector and has since expanded into professional services. Its client roster now includes PwC, Deloitte — for specialist M&A agents — and, most recently, BDO.

The commercial trajectory has matched the adoption curve. The company crossed the $1bn (£748m) valuation mark and raised more than $100m within its first 18 months of operation — roughly $5.6m in funding secured per month since launch.

Model ML has also borrowed a habit from its clients: it embeds staff in clients' offices to, in its words, "understand the business as a whole".

"We're not just a vendor, but we're really an extension of their team," Sehgal said. In a risk-averse industry like consultancy, he added, "like in any relationship, whether that's personal or professional, trust is earned over time."

The valuation is secured. The adoption is measurable. The return, for now, is still a management problem.