Inside our free AI-readiness audit: how we score whether AI is worth it for you

Our free AI-readiness audit asks you nine questions, takes about three minutes, and emails you a scored report. Somewhere between a quarter and half of the people who take it get told, in plain language, not to hire us yet.

That’s by design. This post opens up the scoring machinery — the five dimensions, the exact point values, the four result bands, and why the rubric is built to say “not yet” — so you can decide whether the audit is worth your three minutes, and so you know how to read the number when it lands in your inbox.

What the audit actually is

Three facts about the mechanics matter more than anything else:

• Scoring is deterministic. The same answers always produce the same score. There’s no salesperson adjusting the number, and no AI model deciding whether you “feel” ready. The rubric is fixed code.
• The rubric is the same for everyone. A solo bookkeeper in Lincoln and a 40-person contractor in Folsom get scored on identical criteria. Your score is a measurement, not a pitch.
• The narrative is written around the scores, not the other way. Claude drafts the plain-English sections of your report — what your bottleneck maps to, what your next 90 days should look like — but the numbers are locked before it sees a word of your answers. It explains the score; it can’t move it.

You can see a full sample report before answering anything. It’s the same document you’d receive, with the same honesty settings.

The five dimensions

Each dimension scores 0–20, for a 0–100 total. Here’s what each one measures and why it earns the points it does.

1. Repetitive workload (0–20)

The first question is the one most vendors skip: how many hours per week does your team spend on genuinely repetitive work?

• Under 5 hours/week scores 6. Automation ROI is thin at this volume. A focused build runs $8K–$25K; saving four hours a week doesn’t recover that in any reasonable window. Team AI tools at $25–$30/user/month are probably all you need.
• 5–15 hours/week scores 12. This is the threshold where a focused automation starts paying for itself within the first year.
• 15–30 hours/week scores 17. Squarely in payback territory — typically 6–12 months to recoup a build.
• 30+ hours/week scores 20. At loaded labor cost this is likely $50K+/year of recoverable time. Strong case.

This dimension is the anchor because the payback math is unforgiving: if the labor isn’t there, no amount of clever engineering makes the project pay.

2. Data accessibility (0–20)

Where your business data lives predicts build cost better than almost any other single fact.

• Cloud systems with APIs score 20. Best case. Automations plug straight in with no data migration.
• Industry-specific software scores 14. Data is usually trapped inside, but most platforms expose exports or APIs we can build against.
• Spreadsheets score 12. Automatable, but fragile — expect part of any build to formalize them into something with structure.
• A mix of systems scores 10. The first job becomes picking a system of record. Automating against fragmented data multiplies cost.
• Paper and email score 5. This is where automations go to die. Digitizing intake is the prerequisite, and it’s often a project in itself.

Notice that a low score here isn’t a verdict on your business — plenty of profitable Placer County businesses run on paper. It’s a statement about sequencing: fix the data first, automate second.

3. Team AI adoption (0–20)

• No AI use yet scores 8. Not a penalty — but the right first move is $25–$30/seat tools so the team builds intuition before you commission anything custom.
• Experimented a bit scores 12. The gap between trying ChatGPT and trusting a production workflow is exactly what a focused build closes.
• Individuals rely on it daily scores 16. Adoption risk is low; the opportunity is moving from personal productivity to process automation.
• Team-level use scores 20. Change management is already done. Builds land fastest here.

This dimension exists because the most common way automation projects fail isn’t technical — it’s a team that routes around the new system and keeps doing the work the old way. If nobody on staff trusts AI output yet, a $20K build is an expensive way to find that out.

4. Scale of return (0–20)

Team size, because the same broken process costs more at every additional seat.

• Solo scores 8. Real gains from AI tools, but a custom build only pencils out if it directly unlocks revenue.
• 2–5 people score 14. One automated workflow often returns a full headcount-day per week.
• 6–15 people score 18. Repetitive work compounds at this size; one workflow automated well typically pays back in under a year.
• 16+ people score 20. The same process is being executed dozens of times a day. ROI scales with every seat — though past 50 people, so does integration surface.

5. Problem clarity (0–20)

The last question is open-ended: describe the one task you’d automate tomorrow. We score the specificity of the answer.

• A specific, detailed description scores 20. “Re-typing job details from customer emails into our estimating spreadsheet, about 20 a week” is most of a scoping document already.
• Naming the bottleneck without much detail scores 16. You know where it hurts; the review call sharpens it into measurable scope.
• A fuzzy answer scores 10. That’s normal — but it means the first deliverable isn’t software, it’s naming the problem.

This is the dimension people underrate. In 25+ years and 187+ shipped projects, the single best predictor we’ve seen of an automation project going well is whether the owner can describe the workflow to a stranger in two sentences. If it’s too vague to describe, it’s too vague to automate.

The four bands

The five scores sum to a 0–100 total, which lands you in one of four bands:

Two of those four bands tell you not to buy anything from us. That ratio is the point.

Why “not yet” is the most useful result

A readiness audit that always says “ready” is a lead form with extra steps. Ours will talk you out of a build for the same reason our pricing is flat-bid rather than hourly: misscoped projects are bad for both sides. A $20K automation sold to a business with paper intake and an unmeasured workload doesn’t produce a payback — it produces a resentful client and a dead system.

The economics work because the prescription for a low score is genuinely cheap. If you land in “Foundations first,” the report’s advice costs $25–$30 per user per month and two weeks of honestly tracking your repetitive hours — and most businesses that do those two things are build-ready within a quarter. When they come back, the scoping conversation takes half as long because the audit already did the triage. We’d rather have that client in six months than the wrong project today.

How to read your result

When the report lands, resist the urge to fixate on the total. The per-dimension breakdown is where the useful information is:

• A high total with one low dimension tells you the sequencing. High workload and team adoption but a 5 on data accessibility means the build is worth doing — after (or alongside) getting intake off paper.
• A middling total spread evenly usually means the honest move is seat-level tools now and a re-audit in a quarter. Nothing is broken; nothing is ready either.
• A low workload score with everything else high is the trap case. Enthusiastic team, clean data, no volume. This is where businesses overbuy. Don’t — use the $30/month tools and enjoy them.
• A high workload score with everything else low means the money is real but the foundations aren’t. That’s the “Foundations first” 90-day checklist doing its job.

The score also carries forward: it’s the same rubric we use in paid discovery, so if you do book a review call, the report is the agenda. Nothing gets re-asked.

What the audit can’t tell you

Honest scope for the tool itself: the audit is triage, not scoping. Nine questions can tell you whether a build is worth pursuing and at what tier. They can’t tell you the exact price within our $8K–$25K range, which of your systems we’d integrate first, or what happens when the automation hits an edge case at 2 a.m. — that takes a real conversation and a look at your actual systems, which is what our process covers in discovery.

The audit also doesn’t evaluate failure cost — what happens when the automation gets something wrong. A workflow where a human reviews the output before it ships is a very different build from one that acts autonomously, and that distinction changes both design and price. It comes out in discovery, not in a form.

If you already know your workload, your data situation, and your bottleneck, you can skip the audit entirely and talk to us directly. The audit exists for the more common case: you suspect there’s something automatable in your business but you’re not sure whether it’s a $30/month problem or a $15K problem, and you’d like an honest answer before anyone gets on a call.

Take it or don’t — but measure something

If you take one thing from this post without ever touching the audit, take the rubric: hours of repetitive work per week, where the data lives, whether your team already trusts AI tools, how many people repeat the process, and whether you can describe the bottleneck in two sentences. Score yourself honestly on those five and you’ll land within a band of what our form would tell you.

If you’d rather have it in writing — with the per-dimension breakdown, the tier recommendation, and the 90-day plan — the audit is free and takes three minutes. It arrives by email, there’s no sales call attached, and if the honest answer is “not yet,” that’s exactly what it will say.

Related reading: how much AI automation actually costs · what Claude 4 means for a small business · our AI services · AI automation in Folsom