An innovation consultancy sells something that resists being a product: judgment. A senior strategist looks at a messy situation — a market shift, a workforce disruption, a future that has not arrived — and reasons toward a recommendation that is right for this client, this constraint, this moment. That reasoning is the value. It is also the bottleneck. It does not scale, it walks out the door at 6 p.m., and it costs the same whether the underlying work is novel or the fourth time this quarter someone has run the same landscape scan.
So the pull toward applied AI is obvious and correct: there is enormous repeatable scaffolding around every engagement — research, synthesis, framing, deck-building — that eats senior hours without using senior judgment. Productize that and you free your best people for the work only they can do. But there is a cliff right next to that opportunity. Productize one step too far and you start automating the judgment itself, shipping clients a generic answer dressed up as strategy. That does not scale your value. It commoditizes it, and a generic answer is exactly what a client could have gotten without you.
The problem: the scaffolding looks like the work, but it isn't
From the outside, a great deal of consulting looks automatable, because the artifacts look similar engagement to engagement. The landscape scan, the trends synthesis, the framework, the readout — they rhyme. It is tempting to conclude the whole thing is a template a model can fill in.
It is not. The artifacts rhyme; the judgment underneath them is bespoke. The trends scan is repeatable. Deciding which two of forty trends actually threaten this client and why is not. Confuse the artifact for the insight and you build a machine that produces confident, plausible, generic strategy — the precise thing that destroys a consultancy's reason to exist. The failure mode is not a slow deliverable. It is a client who realizes your "strategy" is something they could have prompted themselves.
Why it is stuck: firms automate everything or nothing
Most firms land in one of two ditches. Some refuse to touch AI, and their seniors keep burning hours on research a model could have drafted in an afternoon — so the work is expensive, slow, and the best people are exhausted on low-judgment tasks. Others go all in, wire up a generate button, and quietly let the model's plausible synthesis stand in for thinking — so the work scales and hollows out at the same time.
The real work is sorting which parts of an engagement are inference over information — gather, summarize, structure, draft — and which are genuine judgment under uncertainty. The first is exactly what modern AI is good at. The second is exactly what it should not own, because a fluent recommendation it cannot actually justify is more dangerous than a blank page.
The path: productize the scaffold, protect the judgment
The firms that productize without commoditizing will build on a few principles:
- Keep a human in the loop where being wrong is costly. The recommendation, the framing of the problem, the call about what actually matters for this client — those stay with the strategist. AI assembles the evidence and drafts the connective tissue; the senior decides what it means. The deliverable is faster to produce and just as much yours.
- Separate the rules from the judgment. Your method — the frameworks, the research process, the way you structure an engagement — is repeatable scaffolding you can encode and reuse. The interpretation that runs through it is the judgment layer. Keep them distinct so you can scale the first without ever letting the second leak into a model.
- Start where judgment is expensive and repetitive. First-pass research, source synthesis, and turning raw findings into a structured draft are the highest-leverage builds: senior-priced work that uses none of the senior insight. Start there, not at the recommendation.
- Earn trust with explainability. When a model drafts a synthesis, your strategist — and eventually your client — needs to see what it drew on and why. "This pattern, from these five sources, with this gap" lets a senior verify and build on it. An opaque, confident summary just moves the risk downstream into a client deck.
Productizing a consultancy is not a transformation project that turns advisors into prompt operators. It is a focused question: which part of your engagement is most expensive and most repetitive yet uses the least judgment right now, and what is the smallest system that hands that work to a model while keeping your people on the part clients actually pay for? That is a two-day conversation before it is a roadmap.
The asset is the judgment. Everything around it is scaffolding, and scaffolding is exactly what applied AI should carry. The firms that win will not be the ones that automated the most. They will be the ones that automated the scaffold so completely that their best thinking — the irreplaceable part — is all anyone is paying for.
Black Flag Design builds applied-AI products for decisions that can't afford to be wrong. If this is your world, spend two days with us — we call it a Foundation Sprint.
