AI Consulting for Professional Services

On the current contract, yes — and pretending otherwise is why so many professional services firms quietly stall their AI projects without ever saying so out loud. If a deliverable takes 40 hours and AI makes it 20, and you bill hourly, you have just cut that engagement’s revenue in half while your costs stayed roughly the same. Everyone in the firm can do that arithmetic, which is why enthusiasm evaporates around the middle management layer. But the framing is wrong in an important way: the choice is not whether the compression happens, it is whether you do it deliberately or a competitor does it to you. Once a firm down the road delivers the same scope in half the time, your 40-hour estimate stops being a price and starts being a liability in a tender. The firms that come out ahead are the ones who move the pricing conversation ahead of the capability, rather than discovering the problem when they lose a bid. That is a business model decision, not a technology decision, and it belongs with the partners rather than with IT.

Your incentives invert, and mostly for the better. Under hourly billing, efficiency is a cost to the firm and the client bears the risk of your inefficiency — which is a strange arrangement that persists because everyone is used to it. Under fixed pricing, efficiency is margin, the client gets certainty, and you carry the delivery risk. That last part is the catch, and it is where firms get hurt: fixed pricing on work you cannot scope accurately is just a way of donating money. The prerequisite is knowing what work actually costs you, which requires real historical data on similar engagements — and most firms have that data and have never mined it. So the honest sequence is: understand your own delivery economics first, move pricing on the engagement types you can scope reliably, and keep hourly for genuinely open-ended work. Firms that flip everything at once discover which half of their work they never understood.

In the work that surrounds the expertise rather than the expertise itself. Proposal and tender response drafting, where firms rewrite substantially the same content endlessly. Knowledge retrieval — finding the report your firm already wrote on this exact question three years ago, instead of starting again. Research and first-pass synthesis. Meeting notes and status reporting. Time capture, which everyone does badly and which is pure recovered margin. Document formatting and quality checking. Data preparation before the analysis. None of it is your intellectual property, all of it is high volume, and it is where the non-billable hours go. The pattern that holds across every professional services firm we work with: the expertise is 20% of the elapsed time and 100% of the value, and the other 80% is the automatable part. Attacking the 80% raises your realisation without touching what clients pay for.

This is the most serious strategic risk on this page and it deserves a straight answer: possibly, if you are careless. Professional services firms have always trained seniors by having juniors do the grinding work — the research, the first drafts, the data assembly. That work is exactly what AI compresses first. Remove it thoughtlessly and you save money for three years and then discover you have no mid-levels, because nobody learned by doing. The firms thinking clearly about this are changing what juniors do rather than how many they hire: putting them on review, judgement and client exposure earlier, which is arguably better training than the grind ever was. But it does not happen by accident, and a firm that treats AI purely as a cost reduction on juniors is making a decision about its 2032 partnership without noticing. It is worth naming explicitly in the strategy conversation rather than discovering it later.

Not until it works inside your own firm. This is the most common overreach we see, particularly among management consultancies and agencies: AI advisory appears on the website before anything has been deployed internally, which means recommending tools nobody has used on problems nobody has diagnosed. Clients can tell, and the ones who cannot tell are the ones you least want. The internal project is the credential — it is also the only way to learn what actually goes wrong, which is the part clients pay for. Once you have compressed your own proposal process or your own knowledge retrieval, you have a genuinely earned story including the parts that failed, and the second conversation is much easier than the first. The sequence is: fix your own house, measure it honestly, then sell what you learned.

A free initial consultation to start — a conversation about your delivery economics, your realisation and where the non-billable hours actually go. Not a demo. If there is a real opportunity, the next step is generally the AI Opportunity Audit at around $3,000: a written map of your workflows, the systems and data behind them, an honest read on which engagement types you could scope well enough to price differently, and a ranked shortlist with realistic effort and payback. Yours to keep either way. Where it stacks up, we build and ship the first workflow into production rather than handing over a report — which, given what this page is about, is a distinction we take seriously. The audit generally earns its fee at around 20 or more staff. And yes, we are aware of the irony of a consultancy telling other consultancies that the deck is not the deliverable.