How AI is changing operations consulting.

Five places, in roughly the order they've matured.

1. Tacit knowledge extraction.

This is the change that mattered most. Two years ago, getting the operational judgement out of a senior person's head was a multi-week interview exercise. Three or four sessions with a consultant, transcribed by a junior associate, structured by a senior associate, validated back with the operator, refined, validated again. Expensive, slow, and shallow because human interviewers couldn't process or cross-reference fast enough to chase the second-order patterns.

Today, an AI-structured interview anchored to real recent work surfaces the same content in 90 minutes and catches more of the second-order reasoning than the human version did. (We covered the specific methodology in AI-Structured Interviews.) The economics are unrecognisable. A piece of work that used to cost five figures per knowledge-holder now costs a fraction of that, and the output is better. This is the one transformation that has genuinely changed which kinds of operational engagements are possible.

2. Drift detection between documentation and reality.

The traditional way to keep a Playbook current was a quarterly review meeting. Slow, low-fidelity, often skipped. The contemporary way is continuous AI-assisted comparison between what the documentation says should be happening and what the project management tool, the field service software, and the financial system actually show happening. Where the two diverge, a maintainer gets a flag with a short list of specific deviations, decides what's an improvement to roll into the Playbook and what's a mistake to retrain against, and edits within 48 hours.

This sounds incremental and isn't. Drift was the failure mode that killed most Playbook implementations. The maintenance burden was too high, the signal was too noisy, the cadence was too slow. AI makes drift detection cheap enough to run continuously, which transforms documentation from a snapshot into a living asset.

3. Decision rule synthesis across operators.

When you interview five senior estimators about the same kind of decision, you get five overlapping but slightly different versions of the rule. Two years ago, synthesising them into a single defensible rule with edge cases mapped to the variations was hand-craft work. Today, the synthesis is mostly AI-assisted: surface the common pattern, identify where the operators diverged, flag for the senior team to discuss the divergence, then produce a unified decision document that captures the agreed-upon pattern. (We covered this in Documenting Decisions, Not Just Steps.)

This isn't replacing judgement. It's accelerating the synthesis step so the leadership team can spend their time on the strategic questions, not the editorial ones.

4. First-draft documentation.

A consultant doesn't write a procedure from scratch anymore. They write a structured prompt with the captured material from the interviews, generate a first draft, review it for accuracy, edit it for tone and team-appropriate language, and ship it for validation. The consultant's time per procedure has dropped substantially. The output quality, after edit, has gone up, because the consultant is spending their effort on judgement and edit rather than initial drafting.

This is the most visible AI change, and also the one that's the least transformative. It changes the cost structure of producing documentation. It doesn't change whether the documentation will be useful.

5. Insights report generation.

Fast diagnostic turnaround that used to be logistically impossible is now routine. The synthesis required to produce a useful operational diagnostic from a short conversation once took a senior consultant most of a week. Today it takes a couple of hours of editing on top of an AI-structured first draft, and the result ends up sharper, because AI is better at cross-referencing a conversation against pattern libraries than a person is. Work that used to gate a senior consultant's calendar now scales.

This is one of the changes that's most visible to prospects, because it's the change that touches the sales-process experience. It's also one of the ones consultants are most likely to over-claim about. The reality is that the AI does the heavy first pass, and the senior consultant is the reason the report is actually useful.

Five places where the work looks roughly the same as it did before. These are the things to listen for when a consulting practice claims AI has transformed everything.

1. Reading the room in a leadership conversation.

When the COO says "we've got really strong systems already" and the founder, a beat later, says "but you know, there's always room for improvement," there's a conversation happening underneath those two sentences that no AI is going to surface. The body language. The momentary glance between the two of them. The way the topic gets routed elsewhere ten seconds later. Reading that, and knowing how to respond to it, is consulting work that hasn't changed in a generation.

2. Calibrating change management to the team.

The same technical Playbook implementation lands beautifully in one team and gets quietly resisted in another. The difference is the team, the culture, the leader's relationship to the senior operators, the recent history of "transformation initiatives." Calibrating the engagement to navigate those dynamics is judgement work. AI can suggest options. It can't sense which one will actually work in your specific business.

3. Knowing which problems leadership is actually willing to solve.

Most operational problems in a service business have a long history. Past attempts. Political associations. People who lost arguments years ago and still resent them. A consultant who walks in without sensing which problems are politically live and which are theoretically live ends up scoping engagements that are going to die when they hit the first hard meeting. This is the most under-discussed skill in consulting and the one AI is least able to help with.

4. Spotting the unsaid in interviews.

An AI-structured interview is excellent at probing what the operator can articulate. It's not yet great at probing what the operator can't articulate but is hinting at. The skilled human interviewer asks the question that lands at the edge of the operator's awareness, the question that makes them pause and say "huh, I've never thought about it that way." That's where the deepest insights come from, and it's still human work.

5. Building trust with skeptical operators.

Senior service-business operators have, on average, been through more failed transformation efforts than is healthy. They walk into the first session with a defensive posture and a "let's see if this consultant is actually any different" question in their head. Earning their trust is a human-to-human exercise. It happens in the first ten minutes. AI doesn't help. (Sometimes it actively hurts, when the consultant leans too heavily on the AI in front of the team and signals that the work is being outsourced.)

The shape of a good operations engagement has changed in two specific ways.

Engagements are faster. The 16-week Phase 1 codification engagement we wrote about earlier in this series, covering Insights, Design, Capture, and Codify on your core processes, used to be a 24- to 30-week piece of work. Capture and Codify are where AI compresses the timeline most. Insights and Design are still largely human work and haven't moved. The Phase 2 deployment (Activate, Amplify, Refine, Oversight) follows on its own runway and is also lighter than it used to be.

Engagements are denser. Because the AI handles the volume work (first-draft documentation, drift detection, synthesis), the consultant's hours go almost entirely to the high-judgement work: leadership conversations, change management, sensing what the team is and isn't ready for. The client gets more senior time per engagement hour. The output is sharper. The cost-per-outcome has come down even though the consultant's effective hourly value has gone up.

This is the genuine shift. It's not "the consulting work has been automated." It's "the consulting work has been redistributed so the human time goes where humans matter."

If you're evaluating a consultancy that claims "AI-powered" in 2026, four questions surface whether the claim is substantive.

AI has changed operations consulting in a way that's larger than the average industry shift and smaller than the marketing copy suggests. The work that's most labour-intensive in volume, extraction, drafting, drift detection, synthesis, has compressed in time and cost. The work that's hardest to do well, judgement, sensing, change management, relationship-building, has stayed exactly where it was.

The implication for service-business owners is that you can now get substantially more operations consulting for the same budget than you could two years ago, but only from practices that have actually integrated AI into the work, not just into the website. The way to tell the difference is to ask the four questions above and listen for stage-specific answers.

The implication for us, doing the work, is that the role of the senior consultant has gotten more interesting, not less. Less time on volume, more time on judgement. Less time on documents, more time on what the documents are for. The AI is doing more of the work that used to fill the week. We're doing more of the work that always mattered.