KPMG consulting head says clients want outcomes but still compare hourly rates

Svilena Tzekova runs consulting globally for KPMG, and she has a blunt view of the business she inherited. Selling advice used to earn the firm a lot of money, she told CNBC this week. That part of the business, in her words, is "literally nonexistent anymore." Clients can now get quickly to the level of advice that firms once spent days or months producing, because the knowledge behind it has become much easier to reach.

Her next point matters more to anyone who buys consulting. Clients increasingly say they want to pay for measurable outcomes, Tzekova said. Their procurement departments still line up competing firms by how many people each one assigns to a project and what those people cost per hour.

So the billable hour is under real pressure, and the spreadsheet that keeps it alive is still on the buyer's side. That puts a sourcing director, and the business owner who signs the statement of work, inside a pricing change that is usually told as a story about the firms.

Why an outcome price is hard to write

Tzekova laid out the cost problem. A model built only on people never had to absorb the expenses AI brings: the price of computing power, plus provider fees that rise with the volume of text a model processes. Those sit alongside the traditional cost of employing consultants, and every one of them has to be reflected in a price. She said it is very difficult to reach an outcome-based model that makes commercial sense. Firms have not yet solved the equation, and that will take time.

Under hourly billing, the firm's cost and the client's bill rise and fall together. More hours, bigger invoice. A fixed outcome price breaks that link. The firm takes the risk that a deliverable needs more rework, more people or more model usage than it planned for. Usage fees are the awkward part, because they grow with the amount of work the software does, and a rate card built on people can't see them at all.

For buyers, this means a fixed fee always contains an assumption about how much of the work AI will do. Two bids can quote similar hourly rates and still rest on very different assumptions. For organizations whose sourcing teams give hourly rate the heaviest weight, the comparison could be measuring the part of the cost that is shrinking.

Elixirr's case for never billing by the hour

Some smaller firms say they started out where the large ones are now heading. Elixirr, a London-based challenger consultancy, has never billed on a time-and-materials basis, Emiko Caerlewy-Smith, its head of investor relations, told CNBC. The firm charges for fixed deliverables, outcomes and value, and she believes a client shouldn't care how long the inputs behind a result take to produce. In her view, AI reinforces a market position Elixirr already held.

The type of work is shifting as well. Clients now want consultancies to help build solutions as well as recommend them, she said, and Elixirr recently helped a global insurer build an AI-based motor insurance platform. That suggests one reason outcome pricing is easier on build work: a working platform gives both sides something to define, test and accept, which a recommendation deck rarely does.

Caerlewy-Smith also gave a useful number. Elixirr spent six months getting its data into a format its AI models could fully use. She wondered aloud how long the same job would take at an organization with tens of thousands of employees, and argued that smaller firms can adapt faster. For a buyer, six months is a concrete benchmark. A firm's AI-assisted delivery is only as ready as its own internal data, and asking where a bidder stands on that is a fair question.

There is a counterweight, and it is about where the money goes. Large firms generated 77% of UK consulting fee income in 2025, according to the Management Consultancies Association's Annual Industry Report 2026, with medium-sized firms at 20% and small firms at 3%. Most consulting spend in that market therefore sits with the large firms now modernizing, a group that includes the Big Four of Deloitte, PwC, EY and KPMG. For most buyers, the pricing model that counts is whichever one those firms settle on.

More graduates hired, less routine work to train them on

The second pressure point is the graduate ladder. Consulting hires thousands of graduates every year, and their careers traditionally start with research, data analysis and presentation building. Those are the tasks AI can now largely automate. Even so, the MCA's 2025 figures, compiled by Oxford Economics from more than 80 firms, show hiring still growing.

Taken together, the figures show the bottom of the ladder getting bigger in 2025. The 37% experienced-hire share points to a second trend running alongside it. The MCA ties that share to more complex client demand and the need for specialists in AI, data, digital change and cyber security. Firms appear to be hiring expertise at the top while they rethink what new joiners do.

Audit gives the clearest example. Junior auditors once spent much of their time on manual "ticking and tying", checking figures against source documents, Paul Goodhew told CNBC. He is EY's global assurance innovation and digital leader and is based in London. AI can now handle the first preparation of that work for the audit professional. EY is also creating new roles: data delivery professionals who get data ready for audit teams, product managers who help build technology platforms, and specialists who understand the technology strategies and risks of clients. For a CFO or audit committee, next year's engagement team may include roles the old org chart didn't have.

Claire Trachet, a former Deutsche Bank M&A banker who founded the London advisory firm Trachet, sees the catch. AI helps juniors finish tasks faster, she said, but it isn't clear how they build judgment if they never do the groundwork. Some juniors are also asking why they should put in long hours on work a tool finishes with one click. Tzekova's answer is to keep hiring graduates and hand them more responsibility earlier. She said some of KPMG's graduates and junior staff are already among those doing the most to drive the firm's new way of working. Caerlewy-Smith put the stakes in plainer terms: firms need to move juniors onto other work quickly, or the economics of the pyramid stop working.

Outside the industry, Brookings Metro senior fellow Molly Kinder argued in a January op-ed that the medical residency model offers a template for keeping entry-level jobs as AI takes over routine tasks: structured, supervised responsibility in place of apprenticeship by grunt work. The MCA puts current formal training at an average of seven days per consultant in 2025, and 7.8 days at smaller firms.

Rewriting the consulting scorecard

The practical lesson for buyers runs back through Tzekova's procurement point. An evaluation sheet that scores bids on hourly rates and team size will tend to favor the time-based model its own business owners say they want to leave. Changing that is a sourcing decision, and it can happen in the next request for proposal rather than waiting for the firms to agree on a model.

Three questions, offered here as an editorial suggestion and not an industry standard, fit the facts these firms describe:

1. How is AI usage priced: bundled into the fixed fee, passed through at cost, or capped?
2. What exactly is the outcome, and how will both sides measure it at sign-off?
3. Who does the work a junior used to do, a person or a tool, and who reviews it?

Hourly rates are not going away on their own. Tzekova herself says procurement teams still compare proposals on headcount and hourly rates, and that arriving at a commercially sensible outcome-based pricing model is very difficult. Firms are still working out the equation, she said, and that will take some time.

The MCA expects UK consulting revenue to grow 6% in 2026 and 8% in 2027, and the demand it cites is for digital transformation, AI and cyber security advice. The next annual report will show whether graduate intake keeps climbing alongside experienced hires, which would be the first hard evidence of whether the redesigned junior role is holding up.