Claude does it in half a day

Monday morning, 9:12 am, at your client's headquarters. You sacrificed your whole weekend for this proposal: four man-days, a solid methodology, a price you know how to defend. Across the table, Nathalie, the procurement director, scrolls through your slides without a word. You know this silence: it usually comes before a question about the timeline. But this time, she leans back in her chair and smiles: "You're not seriously going to sell us four man-days? Claude does it in half a day." Caught off guard, you open your mouth, but your mind goes blank. Not a word comes out. A whole weekend of expertise, swept away by a chatbot's afternoon. 

No, this is not an episode of Black Mirror, but a situation already encountered in sales  meetings. It sums up where consulting stands in 2026: clients no longer wonder whether consultants use AI. They assume they do, and they factor it into the price.

So, is the future of consulting at risk?

To find out, we interviewed Xan Sanchez, senior consultant at Ressource Consulting, a Paris firm of about twenty people specialising in indirect procurement and cost optimisation.

Xan is the firm's AI referent. He took on the role somewhat by force of circumstance. Having worked in digital transformation consulting, he quickly found himself answering all his colleagues' questions.

As for AI's impact on his profession, his view is nuanced.

The man-day put to the test by AI

In survival training, the rule of three sets vital priorities: three minutes without air, three days without water, three weeks without food. Consultants have their own version, just as sacred: three key messages, three recommendations, three bullet points per slide. Even our interviewee announced three reasons for his interest in AI, "because there always have to be three", before realising he only had two. Nobody is perfect, not even a consultant.

For a while, AI productivity gains were not reflected in consulting prices: work went faster, but rates stayed the same. Those days are over. Clients now expect a service that is already "AI-boosted" and are reluctant to pay for the hours it used to take.

The shift is affecting the whole industry, up to the largest firms: clients are questioning time-based billing and asking for outcome-based fees (Valliere, 2026).

Within our interviewee's firm, AI operates at different levels: "commodity" conversational assistants, and Copilot licences being rolled out across Word, Excel and PowerPoint, with access to the employees' SharePoint.

But the firm goes further, with a project tailored to its line of work. Part of its fees depends on the savings achieved for the client. Yet the diagnostic phase, essential but lengthy, costs money without necessarily bringing any back. Hence the tool it is developing with a French AI institute to speed up its fleet and mobility assessments: less time spent collecting and processing data, more time to support the client.

More broadly, AI allows a firm of about twenty people to develop its own automation tools without having to hire a whole team of developers, as large firms do.

Has the consultant become redundant, then?

Not so fast. At our interviewee's firm, consultants set the approach themselves and rely on AI only to build formulas, never to produce a turnkey deliverable. The tools are impressive, though: with a few well-prompted iterations, Copilot in Excel now produces models that would take hours to build by hand. But tone and form remain a blind spot: you don't write the same email to a director, a supplier or a candidate, for instance, and AI rarely senses the nuance. It is also far from infallible. It is even said to hallucinate! On legal or tax questions, it may cite a repealed article, or draw on supposedly expert websites that were themselves written by artificial intelligence.

And even when AI is right, everything happens after the algorithm's answer. A client can get a savings estimate in ten seconds without leaving their desk, but will they trust that figure enough to launch an action plan? Implementation still means negotiating and convincing humans. "Maybe one day all our computers will talk to each other, and we'll be by the pool with a glass of rosé," jokes Xan. We're not there yet.

What about juniors?

Another touchy question is the future of juniors. At large firms, entry-level hiring has fallen by 10% since its 2023 peak. Since 2025, AI roles have outnumbered junior roles there (Revelio Labs, 2026). Yet, as Tarki and Raczynski (2026) point out, these firms turn only one or two juniors in a hundred into partners: smaller intakes could dry up the pool of future leaders.

At Xan's firm, there is no question of cutting junior roles, but rather of changing what is expected of them: among technical skills, Excel proficiency, long a must, is gradually giving way to AI proficiency.

More than an asset, it is becoming a prerequisite: what makes the difference is the ability to write a precise request, challenge the output and rework it. On paper, an AI-savvy junior can already hold their own against a senior.

In an experiment, BCG consultants using GPT-4 worked about 25% faster, with over 40% higher quality, but made more mistakes on tasks beyond AI's capabilities (Dell'Acqua et al., 2023).

In front of a client, however, perspective, experience and quick thinking still make the difference. As our interviewee puts it: "In front of the client, you remain a senior, and the junior remains a junior."

The real danger, then, is not that AI replaces consultants, but that it makes them lazy.

Presenting a paragraph you cannot explain, because you did not really write it, is the surest way to lose a client's trust.

Back to the meeting room. Faced with Nathalie, the procurement director pitting you against Claude or ChatGPT, should you close your laptop and leave the room? Certainly not.

The question is no longer whether AI will replace consultants, but whether they will still be able to explain what they sell. Consultants produce less and less of the analysis themselves: they steer it and challenge it, and clients will pay for their judgment and ability to take action rather than for time spent.

Before we left, Xan Sanchez slipped in a joke: "You're not writing your article with AI, I hope?" Rest assured, our brains worked hard.

1. Dell'Acqua, F. et al. (2023) : https://www.hbs.edu/faculty/Pages/item.aspx?num=64700
2. Revelio Labs (2026) : https://www.reveliolabs.com/news/business/the-industry-that-advised-disruption-is-being-disrupted
3. Sanchez, X. (2026) Interview by C. Contorino, N. Oudin and C. Collet, 15 September 2026.
4. Tarki, A. and Raczynski, J. (2026) : https://hbr.org/2025/10/how-ai-is-upending-how-consulting-firms-hire-talent
5. Valliere, D. (2026) : https://www.thedrum.com/opinion/time-s-up-for-billable-hours-how-should-firms-price-now

Article written by Camille Contorino, Nicole Oudin and Camille Collet