AI made rented time obvious

If a client can talk you out of your fee by pointing at AI, the fee was never about judgement; it was about hours, and AI just made that visible.

That can sound defensive when it is used to protect a rate card. I mean it as a test.

For a long time, time was the easiest thing to sell because it was the easiest thing to see. Research took days; a first draft took a week. A prototype, competitive review, strategy deck, process map, or requirements document carried visible evidence of effort. Clients could point to the calendar, the number of people involved, and the size of the artefact at the end.

Some of that effort was real and necessary, and some of it was the work. A lot of it was the friction around the work, and that friction quietly became the scarcity clients were paying for.

AI has not removed the value of serious consulting. It has removed the convenient excuse for pricing serious consulting as if keystrokes, slides, research hours, and the first version of an answer were the product.

The first version is cheaper now, and that is the change.

The work was never only the artefact

A client can ask an AI tool for a market scan, a proposed user flow, a set of workshop questions, a competitor summary, or a draft operating model. In many cases, the result will be useful enough to start a conversation.

Clients should be able to get a first pass without paying someone to perform effort in public.

The question starts after the first pass arrives.

Is the market scan using the right frame for this company, or has it given the client a convincing summary of the wrong market? Does the user flow fit the actual business model, operating constraints, and people who will use it? Does the proposed workflow preserve the decision that matters, or has it turned a difficult trade-off into a neat diagram with no owner?

AI can create material quickly and make a bad direction look unusually finished.

I separate three kinds of work that get collapsed into one fee.

  • Execution is getting something made: the draft, the research, the board, the prototype, the workflow, or the automation.
  • Judgement is deciding whether that thing is the right intervention, what it should contain, what it must exclude, and what consequence follows if it is wrong.
  • Ownership is deciding who carries the decision once the consultant has left, who can challenge it, and who accepts the cost when conditions change.

AI has made execution cheaper. It can contribute to judgement when it is given good inputs and a defined decision. It cannot accept responsibility for a trade-off, read a founder’s unstated constraint, or decide that the apparently efficient option will damage trust with a customer, regulator, partner, or team.

Someone still has to do that work.

“AI can do this now” is often partly right

When a client says that, I do not think the right response is to defend the old hours.

Sometimes they should build it themselves.

If a founder or team can use a tool to create the first version of a process, automate a recurring task, or turn scattered notes into a working draft, they should have that capability. A consultant who tries to keep basic tool operation mysterious is protecting a weak position.

I run AI workflow work that includes teaching founders and teams how to set up and run parts of the tooling themselves, not as a free sample designed to make them dependent on a larger engagement, but so they can handle the work that is genuinely repeatable.

That changes the relationship in a useful way.

The client does not have to come back for every minor adjustment. I do not have to pretend that I should own a process their team can run. Both sides can see where the handover ends and where a new problem begins.

The work becomes more honest. They own the tool and the day-to-day operation. I remain useful when the question becomes harder: which workflow deserves attention, what should the system know, what needs human review, how should the team judge whether the output is safe to use, and what should change when the first version fails under real conditions?

That is a more defensible place to create value.

Tool access is not operating capacity

A team can buy the same model, connect the same automation platform, and copy the same prompt structure as any consultant. Those things are increasingly available.

What they cannot download is the accumulated judgement behind the choices.

Why is this workflow worth automating before the other ten? What information is allowed into the system, and what stays out? Which output can move without review, and which one needs a person who understands the consequence? Where is the evidence for a recommendation? What happens when an exception appears? Which owner has the authority to change the rule?

Those questions are not glamorous. They are where real AI work succeeds or fails.

A tool can generate a response in seconds and surface the options, but it cannot tell a UAE founder whether that response will create a customer-support burden, expose sensitive data, weaken an approval process, or create a promise the team cannot keep, nor can it own the commercial and operating consequence of choosing one.

A serious AI engagement starts with the work around the tool: the recurring decision, the inputs, the handoffs, the exceptions, the review point, and the person who will still be responsible on a difficult day.

The tool matters, but it is not the whole intervention.

An AI workflow implementation should establish the right workflow, transfer what the team can operate, and stay available for the decisions that need deeper product, systems, or business judgement rather than make a team more dependent on outside expertise.

Teaching the client is part of the offer

There is a strange assumption in consulting that the safer commercial model is to keep the client unable to do the work without you.

That model can produce recurring revenue and a weak relationship, because the client eventually notices that the consultant’s usefulness depends on their own lack of capability.

I would rather make the boundary visible.

A capable team should own parts of an AI workflow: running an established prompt or automation, maintaining an agreed source, checking a known output, and making routine changes inside a defined rule set. Teaching those things is useful, and it makes the engagement more credible.

The harder questions deserve a more experienced challenge: choosing the first workflow, deciding whether an apparent bottleneck is even an AI problem, reconciling competing business priorities, designing a review model for sensitive work, or deciding when a successful pilot has become an operating risk.

Those are not tasks someone should outsource forever. They are decisions where outside judgement can be valuable precisely because it is not embedded in the team’s daily assumptions.

The goal is a team that can do more of its own work well and knows when the next question requires a different level of attention, rather than dependence on a consultant for routine operation.

Price the decision, not the performance of effort

A useful fee should be explainable without talking about how many hours it took to produce a first draft.

The client should be able to ask: what decision will this work help us make, what evidence will it use, what will we own when it is done, and what becomes easier or less risky because the decision has been made properly?

Those questions make the work inspectable. They also protect the client from paying for theatre.

A fee based only on a consultant knowing how to operate a tool deserves scrutiny. Tool operation is becoming cheaper, and basic execution should become cheaper with it.

When the consultant can frame the right problem, identify the constraints that matter, decide where AI belongs, build a working path, and leave the team with an operating model it can use, the value is tied to the quality of the next decision rather than the number of hours needed to create the first artefact.

AI is forcing consultants to be more explicit about that distinction, and it should.

Price on the judgement that helps a client avoid the wrong build, make the right decision sooner, and leave with more capability than they had before. Everything else was always rented time.

Keep Reading

Related Posts