How it works
From a name and a location to a plan you can act on
The audit is fully automated and built on your market. Here's exactly what happens between your snapshot and your report.
- 01
We build your question set
For your trade and location we assemble 20–40 buyer-intent questions — the things real customers type when they're ready to hire, like “best accountant for contractors near me” or “who's the best accountant for VAT in Leeds”. Note: the free snapshot uses a small set of generic questions only — the full audit adds questions naming your specific specialisms.
- 02
We ask the three major AI assistants
Each question is put to ChatGPT, Claude and Gemini, live. We capture the full answer — every firm named, in the order and language the AI assistant chose. Note: the free snapshot checks ChatGPT only; the full audit covers all three.
- 03
We read who got recommended
We parse each answer for your firm, your competitors, and the descriptive language AI attaches to the winners — the specialisms and phrases that earn a recommendation.
- 04
We score and benchmark
You get a visibility score per platform, a share-of-voice against local rivals, and a read on where national brands cap the local field.
- 05
We write your playbook
A verdict, the competitive picture, a gap analysis, and a ranked action plan. We rewrite the specific website page copy holding you back.
- 06
You get a PDF by email
The full report arrives in your inbox within two hours of purchase. No login, no dashboard, nothing to manage — just the report and the plan.
Full transparency · no black box
The exact questions we ask
Your audit runs 20–30 real buyer-intent questions typically — up to 40 if your trade and location have a lot of specialisms to cover — built from a library tuned to your business and expanded for your town and specialisms. They fall into five types, ordered by how close the buyer is to hiring:
Location + specialism
Highest intent- “accountant specialising in IR35 in Leeds”
- “contractor accountant Leeds”
Location + category
Highest intent- “best accountant for small businesses in Leeds”
- “top accountants for sole traders in Leeds”
Problem-based
Highest intent- “accountant for IR35 in Leeds”
- “R&D tax credits specialist Leeds”
Comparison
High intent- “recommended accountants near Leeds”
- “best rated accountants in Leeds”
General
Research stage- “how to find a good accountant for a freelancer in the UK”
Examples shown for an accountant in Leeds. Other types of professional services firm get a different set. We cap the set at 40 and always guarantee a slot to every question type, so a firm with many specialisms never crowds out the broader searches — and we always lead with the highest-intent questions, because that's where the buyers with money to spend are asking.
The free snapshot samples 6 generic questions from this library, on ChatGPT only — we don't collect your specialisms until the full audit, so the location + specialism and problem-based questions above are exclusive to it.
How the score works
How we calculate your visibility score
One number, defined plainly — so you can check it against the verbatim answers in your report's appendix.
Your visibility score is the share of answers that name your firm.
We count how many answers named you and divide by the number that came back successfully — reported per platform and overall. For example, named in 6 of 24 answers is a visibility score of 25%.
Failures don't count against you
If an AI assistant times out or refuses, that answer is left out of the maths entirely — a technical glitch can never be mistaken for you being invisible.
Share of voice, in context
Of every firm the AI assistants recommended across your questions, we measure what share was you, versus local competitors, versus national brands.
National brands, kept separate
Big national chains are identified against a maintained list and reported on their own, so they don't flatter or distort your local picture.
What the numbers are — and aren't
A point-in-time reading, disclosed in full
No sampling tricks, no black box. Your report states exactly how it was produced — so you can trust the number and know what it does, and doesn't, mean.
Each question asked once
Every question is put to each assistant a single time — one unaided reading, exactly as a buyer would see it. We don't repeat questions to average out a bad run or cherry-pick the best one.
The exact models, on the record
We record which model each assistant was running — ChatGPT (GPT-4o-mini), Claude (Claude Haiku 4.5) and Gemini (Gemini Flash) — and your report names them alongside the date it ran.
A snapshot, not a fixed ranking
AI answers shift over time, so your audit is a point-in-time reading dated to the day it ran. The fixes we recommend target the durable signals that move your visibility whatever the day-to-day wobble.
Two parts of the report are built by reading your website — the technical readability checks and the suggested copy rewrites. If your site's security or bot-protection (such as Cloudflare) declines our automated request, those parts may be limited or left out, and the report tells you where. Your visibility score and the competitive analysis are unaffected, since they don't rely on reading your site. Blocks like this are often intermittent, so re-running later may succeed.
The substance principle
A real report — even if AI never names you
Most firms start out invisible, so a report built only on your own results would be thin. Ours isn't. The depth comes from the market — the dense competitive field across 20–40 questions. We state your own numbers honestly and compactly, then spend the report on what actually helps: who wins, how, and what to copy.
Start with the free snapshot
See where you stand in 30 seconds, then decide on the full audit.