I install an AI operating system inside a service business that makes it easier to find, easier to trust and easier to choose, starting with where it stands among the businesses I have scored in its market.
Drop in your website. Nothing to install and no access to hand over.
Start 126 items · 9 sections What we scoreEvery moment a customer decides against you, marked 0, 1 or 2.
See the checks The part nobody else has Your rank in the fieldFifth out of twenty means having scored the other nineteen.
See the benchmark Published whole TeardownsReal businesses taken apart, every file we produced, unedited.
Read oneNo call required. Results in about five minutes.
Nearly every owner we have scored got sold marketing before anyone told them what was actually broken. Diagnosis first is the whole idea.
Drop in your website. Nothing to install and no access to hand over, because the first pass reads you the way a customer or a search engine does.
Your results from the 126-item scorecard, broken down by section, so you can see which link in the chain is thin instead of guessing at the whole thing.
The exam writes out one fix, the one that moves your score most, not the one that is easiest to sell you.
If you want more than the one fix, we talk on a call about the Diagnostic or an install. Every paid step starts on a call, and nothing is bought online.
These are measured failure rates from one market we scored end to end, twenty roofers in Birmingham, every one run against the same scorecard. Almost none of it is exotic, and that is the point. Nineteen of those twenty had nothing at all catching a lead after hours, which means that tonight, the customer who cannot reach them calls somebody who answers.
A customer decides against you somewhere specific: a page that will not load, a review they cannot find, a form nobody answers. Every one of those moments gets scored, each item marked 0, 1, or 2. The nine sections answer three questions: easy to find (sections 1 to 3, 45 items), easy to trust (sections 7 to 9, 30 items) and easy to choose (sections 4 to 6, 51 items).
SSL, load speed, mobile rendering, whether a browser throws a warning before anyone sees your work.
Google Business Profile completeness, map pack position, citation consistency, service-area pages.
Schema, llms.txt, and what GPTBot, ClaudeBot, and PerplexityBot actually see when someone asks who to call.
Navigation, accessibility basics, and whether the pages that are supposed to sell actually do.
Form length, booking friction, pricing transparency, and every other reason someone leaves without contacting you.
Live chat, CRM connection, call tracking, and whether anything at all catches the lead that arrives at 9pm.
Volume, recency, response quality, and whether any of it is visible where a customer actually lands.
Certifications, license visibility, years in business, and whether a reader could say why you and not the shop down the road.
What you publish, how often, and whether any of it is doing work for you a year later.
The free tools are not bad, they are just scoped to a website. Most of what decides whether a lead becomes a job happens somewhere else.
Anyone with an API key can grade a website against a checklist. Telling you that you sit fifth out of twenty means having scored the other nineteen, and that is the part that cannot be bought.
There are no testimonials on this page, no client logos, and no results claims, because there are no paying clients yet to draw them from. What exists instead is the work, published whole.
Five real service businesses in five different trades, every file we produced for each one, published in full rather than as a summary. The company details are replaced throughout and nothing else was touched.
We run entire map packs through the same 122 checks and publish what comes back. No company is named, because none of them asked to be scored, though the trade, the city and every number is real.
This exam is going to tell you that you are losing jobs, so it had better be right. Three rules make it something you can act on instead of argue with, and you can hold us to all of them.
If a check cannot be assessed on your business, it is excluded from your score rather than counted against you. Scoring unknowns as zeros is how most automated audits manufacture a scary-looking number, and it is the single most common reason an audit is wrong.
Each check records what was observed and where, so a finding is something you can go and verify on your own site rather than a claim you have to take on faith.
The scores are measured from what is on your site and in public data. I do not put a dollar figure on a gap and I do not forecast results, because a precise number about a business I have not worked with would be a guess dressed up as arithmetic.
The Business Health Exam is free. If you want more, the Diagnostic is $2,500 and is credited in full toward either install you sign within 45 days. All prices are one time.
One-time prices·No subscription·Nothing bought online
Not sure where you stand? Start with the free Business Health Exam. Start with the free exam
The full picture of your business, and the 10 businesses closest to you. Credited in full toward either install signed within 45 days.
Six owners, one week of live calls: a 1.5-hour morning call and a 1-hour evening call, with standby between sessions.
Six days on site with one business. Travel is billed at cost and estimated in the proposal.
A 50% deposit reserves the date and the balance is due at the start. You pay for your own Claude subscription and any usage directly. My fee covers the install, not the software. I take one on-site install a month, and six owners per group. Best fit: owner-operated businesses where the owner still makes the decisions, that already run a CRM or scheduling software, and can hand over admin access before the install.
The exam is free and nobody calls you about it unless you ask. Drop in your site and the first pass runs on what is publicly visible, the same way a customer or a search engine sees you.
I built the scoring engine and I run it myself. No account managers, no strategy slides, no "we'll circle back."
My background is in government intelligence operations, which is mostly the discipline of not stating things you cannot support. That is why an unknown on your scorecard stays an unknown instead of quietly becoming a zero, and why you will not find a client logo on this page until there is a client who agreed to be on it.
A free exam that scores your business on a 126-item scorecard across the whole path from someone needing you to a job on your calendar. You get a section breakdown showing which link in the chain is thin, a check of whether AI assistants name your business, one fix written out, and your rank among the businesses I have scored in your market.
A free tool grades your website, which is one of nine sections here. It has nothing to say about whether a lead that lands at 9pm gets answered, whether your reviews are visible where customers actually look, or how you compare to the business across town. It also cannot rank you, because it has not scored anybody else.
They are excluded from your score, never counted as a zero. Scoring an unknown as a failure is the fastest way to make an audit look thorough and be wrong, so when I rank a market I compare only on the checks every business in it could be assessed on, and I publish that number alongside the ranking.
It is $2,500, one time. You get the full findings of my scan on your business, a ranked rollup of your 10 nearest competitors, and one walkthrough call. It is credited in full toward either install if you sign within 45 days. It starts on a call, and nothing is bought online.
The Group Install is $8,000 and the On-Site Install is $20,000, both one time. A 50% deposit reserves the date and the balance is due at the start. You pay for your own Claude subscription and any usage directly. My fee covers the install, not the software.
Owner-operated businesses where the owner still makes the decisions, that already run a CRM or scheduling software, and can hand over admin access before the install. I take one on-site install a month, and six owners per group.
From scoring a whole market at once. Dallas commercial roofing, 20 companies, scored on 20 August 2026, and Oklahoma City plumbing, 34 companies, scored in August 2026, both on the 122-item version of the scorecard in use then. Birmingham roofing, 20 companies. Every business is audited independently against the same scorecard, then compared only on the checks all of them could be assessed on.
Your section breakdown and your one fix, which is most of the value and does not depend on anyone else. The ranking arrives when your market gets run, and naming a city is usually enough to put it in the queue.
Only if you want it. The exam is free and stands on its own. If you want more, we talk on a call about the Diagnostic or an install. If the exam shows your problem sits somewhere I do not work, I will tell you that.
Yes. Five complete builds are published in full and unedited, 192 files and 249,162 words across five businesses in five different trades, with the company details replaced and nothing else touched.
Today is the worst shape your business's health will ever be in, though that is only true if you find out where it actually stands. The first exam is free and takes about five minutes.