Consulting directors advise clients on AI adoption daily but haven't run it in their own practice. Here's the concrete first move to fix that.
Every consulting pitch and its three legged dog at the moment contains some element of Artificial Intelligence. AI readiness assessments, adoption roadmaps, change management frameworks, consultants are selling all of it. But the funny thing is, most of them haven’t run a single AI workflow on their own practice.
You’ve got to practice what you preach, because it matters. A firm that advises clients to pick one specific, bottom-line-impacting use case and start there, then goes back to producing its own reports the same way it did five years ago, has a credibility problem. The good news is the fix is straightforward, and the first move is obvious once you see it.
The work that’s eating your consultants’ time
Lets use an example, say a mid-sized consulting firm. A principal wins a strategy engagement. The team spends the first two weeks pulling together market research, reading industry reports, synthesising competitor analysis, and turning raw notes from stakeholder interviews into structured findings. Then someone drafts the deck sections. Then someone else rewrites them. That cycle, from sources to first draft, takes three to four days of consultant time on a typical engagement. It’s high-volume, repeatable work. The sources change. The structure doesn’t. That’s exactly the kind of task AI handles well.
The pressure this creates is real. Senior consultants end up reviewing work that should have been further along before it reached them. Junior consultants spend the bulk of their week buried in documents rather than building the analytical skills that make them more valuable over time. And the firm’s utilisation numbers don’t reflect what’s actually billable, because too much of the week goes to work that could have been done faster.
There’s also a subtler cost. When a director is preparing for the client workshop on Sunday night because the first draft only landed Friday afternoon, that’s not a resourcing problem. It’s a workflow problem, and it repeats on every engagement. The time lost isn’t dramatic in any single week, but across a full year of engagements it adds up to a meaningful share of capacity that could have gone to billable work or business development.
What the first move actually looks like
The AI ingests your sources, whether that’s uploaded PDFs, web research, interview transcripts, or internal documents, and produces a structured first draft of your findings sections. It organises the material, pulls out the key points, and writes prose your consultants can edit rather than a blank page they have to fill.
Think of it like a very fast research analyst who has read everything and can’t make a judgement call. It surfaces what’s in the sources. Your consultants decide what it means for the client. Nothing goes to the client without a senior eye on it, and that’s how it should stay.
We built this as a pilot that a firm ran on two engagements, it cut the research-and-draft cycle from three days to three hours. The consultants spent the recovered time on analysis and client prep, the work that actually moves the engagement forward and establishes the relationship. By the second engagement using the same workflow, the setup time had dropped too, because the process was already mapped and the AI knew the firm’s preferred report structure.
This is the kind of work that AI agents for professional services are built for: high-volume, structured, and repeatable, with a person staying accountable for every output that leaves the building.
The pilot has to be designed to scale
The principle here, drawn from how we approach first moves with clients, is that the first use case must be viable as a pilot and designed for eventual scale. Research synthesis fits both criteria. It’s contained enough to run on a single engagement without affecting the rest of the practice, and it’s present on every engagement, so the gains compound quickly once the workflow is proven.
The other principle is that the task must move the bottom line. Recovered consultant time on billable work does exactly that. Utilisation goes up. Delivery costs come down. The firm can take on more work with the same team, or deliver the same work with less strain on the people doing it.
One thing it won’t do is fix a poorly scoped engagement. If the brief is vague, the AI will synthesise a lot of material that doesn’t answer the client’s question. The workflow depends on the consultant knowing what they’re looking for before the sources go in. That’s not a limitation to work around later; it’s part of how you set the pilot up correctly from the start.
The AI Strategy work starts here: map the task, set up the workflow, and measure the before and after on a real engagement. That’s the plan, not a vague commitment to do more with AI.
What this means for you
If your firm is advising clients to start with AI, the most credible thing you can do is start yourself. Research synthesis and report first-drafts is the right first move: high volume, measurable, and present on every engagement you run.
Your consultants still own the thinking. The AI clears the first 60 to 70% of the drafting work so they can focus on the 30 to 40% that needs their expertise and their name on it. It won’t write the recommendation. It will make sure the person writing it isn’t spending three days getting to a blank page first.
Take the AI Roadmap Interview and tell it where your team loses the most time. You’ll leave with a concrete plan for where to start, built around your practice.
