Sounds familiar? · Strategy
“We know we should use AI — we just don't know where to start.”
Who feels it: Leadership teams circling the topic for two quarters; managers fielding 'what's our AI plan?' without a defensible answer; companies watching competitors announce AI initiatives and wondering what's real.
The straight answer
The right first AI project is found by arithmetic, not brainstorming: rank your processes by annual cost of manual effort, keep only those that are high-volume, repetitive, digital, and rule-describable, then pick the top survivor that has a willing owner. Four filters, one ranked list, one bounded pilot with a baseline. Companies that start this way get a measured win in weeks; companies that start with 'what's the coolest thing AI can do?' get a demo and a shrug.
Why it happens
The root causes — named honestly
Symptoms get treated and return. These are the structural reasons the problem exists, which is where the fix has to aim.
Starting from the technology's capabilities
'What can AI do?' generates fascinating answers and no decisions. The tractable question is 'what does our operation waste?' — that one has a ranked answer.
Overweighting ambition on attempt one
The transformative moonshot project maximizes risk exactly when organizational trust in AI is at its minimum. First projects should be chosen to build proof, not headlines.
No shared map of process costs
Most companies have never priced their manual processes, so every AI conversation floats. A one-page ranked cost map converts the debate into a decision.
What good looks like
- A ranked list: each manual process with its annual cost and automation fitness
- A first pilot that is high-volume, repetitive, digital, rule-describable — and owned
- A baseline measured before go-live, a success metric agreed before that
- A 90-day path: audit → pilot → verdict → scale-or-stop
The fix
How we get you there — step by step
Take the free readiness score
Eight questions, three minutes, no email required — a first read on where you stand and what to fix before spending anything.
Run the AI Opportunity Audit
1–2 weeks: your processes mapped and priced, opportunities ranked by expected return and risk. The 'where to start' question gets a numerical answer.
Pilot the top survivor
One process, live in under 14 days, baseline first, human oversight built in. Scoped so success is provable and failure is cheap.
Scale on the verdict
Results against baseline decide what's next. The second project inherits the first one's proof — momentum built on evidence, not enthusiasm.
What changes
The measurable difference
3 min
to a readiness read — free, no email required
1–2 weeks
to a ranked, priced opportunity map
90 days
from 'where do we start?' to a measured verdict
Follow-up questions
What people ask next
What makes a process a good first AI candidate?
Four filters: high volume (it happens a lot), repetitive (the same steps each time), digital inputs (data or documents, not paper or phone calls), and rule-describable (a competent person could write down how to do it). Add a willing process owner and a measurable baseline, and you have a pilot.
Should we hire an AI lead before starting?
Not for the first project — you'd be hiring before knowing what the job is. Run the audit and a pilot with a partner first; the results tell you whether you need a full-time hire, a fractional resource, or just trained process owners.
What if the audit shows AI isn't worth it for us yet?
Then you've bought certainty for the cost of a fortnight — and usually a list of cheaper process fixes the audit surfaced along the way. That outcome is rarer than you'd think, but when it's the truth, it's the recommendation.
Is this your situation? Bring it to a call.
A 30–45 minute working session on the actual process — with the person accountable for the outcome, not a sales rep.
No retainers to start · Pilot-first · Founder-accountable