Notes · Essay
Most consulting engagements start with the model and reason outward. The ones that actually help start with the workflow and reason inward.
There’s an epidemic of AI strategy decks right now. They tend to open with a landscape slide showing every AI application from ChatGPT to autonomous driving, argue that the business has to transform or die, and close with a five-year roadmap toward becoming an “AI-first company.” What the deck rarely contains is a specific answer to the question the client actually walked in with: where does AI fit in the work we already do?
This is the reversal. Most AI strategy work starts from the model and reasons outward: what can this new thing do, and how do we deploy it? The work that actually delivers value starts from the workflow and reasons inward: where is our work genuinely stuck, and is there something in this new toolkit that would move it?
A regional nonprofit’s development team spends about twelve hours a week drafting personalized grant follow-ups. They know each grantor’s history, priorities, and past objections. What they don’t have is time. Someone selling AI strategy might sell them a fifty-thousand-dollar CRM integration. What actually helps them is a well-designed prompt template that pulls their existing notes into a coherent draft they can edit in twenty minutes instead of two hours.
The intervention costs a few hundred dollars in setup and gives back eight hours a week. The strategy question wasn’t “how do we transform.” It was “what’s slow, and can we speed it up without losing our voice?”
The right AI strategy engagement usually looks like three things: identifying the two or three workflows where the team is genuinely blocked, testing a handful of interventions against those workflows, and training the people who will do the work to actually use the tools.
That third part is where most engagements fail. A workflow retooled by consultants who then leave is not a workflow that persists. The team either uses the new tool for a week and drifts back to old habits, or uses it wrong and blames the tool.
Team training is the piece most decks skip. If a designer, an analyst, and a project manager each save six hours a week because they learned to prompt well, that’s roughly a full extra person on the team without hiring one. That result compounds. But it only shows up if the training actually happened, and if it addressed the specific tools those people already use.
There’s a smaller but similar version of this for individuals. Someone rethinking their career in 2026 doesn’t need a course on how AI works. They need a personal audit of what they do all day, which parts of that work AI is now doing faster or cheaper, and what to build next. That’s a career strategy conversation, not a technology one, and it looks nothing like a landscape slide.
What specific workflow will look different in six weeks, and how will we know?
If the answer is “we’ll have a roadmap,” walk away. If the answer is “the grant team will be drafting in twenty minutes instead of two hours, and I can measure that in your CRM,” you might be talking to the right person.
The good news for anyone shopping this kind of engagement is that the right consultant is much cheaper than the wrong one. A workflow-focused intervention is measured in weeks and thousands of dollars. A transformation roadmap is measured in quarters and tens of thousands. The correct engagement is smaller than the industry has trained you to expect, which is exactly why it works.
Notes are occasional pieces I write when I have something specific to say. If you want to work through where AI actually fits in your practice, that’s the Generative AI Strategy service. First consultation is free.