AI for management consultants: tools that earn their keep

If you run a boutique or independent consulting firm and are wondering where AI fits, the answer is on the inputs, not the judgement. It earns its keep by speeding research, synthesis and first drafts. The thinking, the client relationship and the sign-off stay unmistakably yours, because that is what a client is paying for.

AI is now part of how the sector works, and it is reshaping consulting's own economics. The MCA Member Survey 2026, researched by Savanta across more than 1,000 consultants and published in January 2026, found 77% of firms had integrated AI into their systems or enabled staff to use AI models, and 78% pointed to digital technology and AI services as a driver of growth in 2026. The useful question is no longer whether to use it, but where it belongs in the work.

Where AI earns its keep

The gains cluster at the front of an engagement, in the document-heavy and research-heavy steps, not at the point where you form a view. The same MCA survey found 76% of firms now use AI to search for information and undertake research, up 9% on the year before. That is the pattern to copy: use it where the work is gathering, sorting and drafting.

Research and desk work. AI is quick at scanning a market, pulling together a first read of a sector, and summarising long reports into something a team can react to. Treat the output as a briefing to check, not a finding to repeat. The facts still need verifying against the primary source.

Synthesis. Turning twenty interview transcripts into themes, or a folder of documents into a structured summary, is exactly the kind of slow, mechanical work AI compresses. It gets you a first cut of the structure. You still decide what actually matters.

Deck and document drafting. A first draft of a findings section, a workshop agenda, or a set of slide headlines is faster with AI in the loop. It removes the blank-page cost. It does not remove the need for a consultant to shape the argument.

Modelling support. AI can draft the scaffolding of a spreadsheet model, write the formulas, and explain what a competitor's numbers imply. Check every calculation before it reaches a client. A confident wrong number is worse than a slow right one.

The commoditisation risk

Here is the honest part. If AI can produce a passable market scan or a tidy summary in minutes, so can your client, and so can every other firm. The work that AI makes cheap is the work that stops being worth much. Leaning on it for the deliverable itself is the fastest way to make your firm look like everyone else.

The MCA's own members see the tension: the same 2026 survey found 52% of firms named rapid AI adoption as the industry's leading challenge, citing higher technology and training costs and uncertainty about what clients will pay for. The value is moving away from producing the analysis and towards judgement, framing and trust, which is the part a model cannot hold. We have written before about why context is the moat: the understanding you hold about a specific client, their politics and their history is what a general model will never have.

This is also why raw speed is the wrong scoreboard. Doing the commodity part faster does not protect a firm. Our note on AI productivity versus business value makes the case that saved time only matters if it goes back into the thinking clients actually pay for.

The disclosure question

Clients are increasingly asking how their advisers use AI, and consulting is no exception. The safe default is straightforward: be able to answer plainly. You do not need to itemise every prompt, but you should be able to say what AI touched, what a person checked, and that no confidential client material went into a tool that could train on it.

Two rules keep you on the right side of this. First, do not put client-identifying or commercially sensitive information into a consumer chatbot that may use it to train future models. Use a business-grade tool with an agreement that keeps your data out of training. Second, never present AI-generated analysis as your firm's considered view without a person having checked and owned it. The client is buying your judgement, and the judgement has to be real.

Keeping the judgement human

The line to hold is simple to state and easy to blur under deadline pressure. AI can prepare the ground. It should not form the recommendation, decide what a client needs to hear, or make the call that carries professional risk. Those are the moments a client is paying a senior person for.

A practical test: if the output would embarrass you or expose the client when it is wrong, a person owns it, end to end. Research inputs, first drafts and formatting can lean on AI. Conclusions, advice and anything client-facing get a human name against them.

A starter workflow

You do not need a programme to begin. You need one bounded task and some discipline about how you use the tool.

• Pick one low-risk, recurring task. Interview synthesis or first-draft research summaries are the natural first step. Prove the time saving on something that never reaches a client unchecked.
• Use a business-grade account. One with a data agreement that keeps client information out of model training, not a free consumer login.
• Write the prompt to stay on the safe side. "Summarise only the attached approved notes into neutral themes, flag gaps, and do not draw conclusions or make recommendations" keeps the tool on the inputs. A loose prompt invites it over the line.
• Keep the human check visible. Every AI-assisted output that informs a client deliverable is read and corrected by a consultant before it goes anywhere.
• Measure against corrections, not speed alone. Track the time saved and the time spent fixing. If the fixing outweighs the saving, the task was the wrong one.

What to do next

Pick one recurring task your team does every week that never reaches the client unchecked. Interview or document synthesis is the usual candidate. Run it through a business-grade AI tool for a fortnight, with a consultant reviewing every output, and measure the time saved against any corrections needed. That single, bounded test teaches you where AI genuinely helps your firm, while keeping the judgement exactly where it belongs.

If it would help to map where AI fits across your engagements, and where a person has to stay in charge, book a call and we will think it through with you. If you are weighing outside help, our note on how to choose an AI consultant is a good place to start.

Common questions

Should management consultants use AI for client work?

Yes, on the inputs. AI is well suited to research, synthesis, first drafts and modelling support, where it saves real time. It is a poor fit for the recommendation itself, the client judgement and anything that carries professional risk, which is what a client is actually paying for. Use it to prepare the ground, not to form the view.

Will AI commoditise consulting?

It commoditises the parts that were already close to a commodity: the market scan, the tidy summary, the standard analysis. That raises the value of the parts it cannot do, which are judgement, framing and trust built on real knowledge of a specific client. Firms that lean on AI for the deliverable will look interchangeable. Firms that use it to free time for sharper thinking will not.

Do we need to tell clients we use AI?

You should be able to explain your use plainly if asked, and clients increasingly do ask. That means being able to say what AI touched, what a person checked, and that no confidential material went into a tool that could train on it. Keeping client-identifying information out of consumer chatbots is the non-negotiable part.