No workshop. No three-week onboarding.
You give us a domain and the questions your customers ask. We go and find out what the models do with them.
- 01 You: 30 seconds. Us: one working day.
The free check
You give us a domain. Our automated suite tests crawler access, detects structured data and renders your key pages without JavaScript. You get a scored report by email with the three things we would fix first.
- 02 30 minutes.
Scoping call
We agree three things: which page templates matter commercially, which competitors to benchmark against, and which customer questions to test. That question set is the most important input you give us, and we help build it from your enquiry data and sales calls.
- 03 Days 1 to 5.
Technical testing
Each AI crawler is tested by name, end to end, and any block is traced to the specific rule and system responsible. We audit structured data across every significant template, and compare raw server responses against rendered output to establish exactly what a non-executing crawler receives.
- 04 Days 4 to 8.
Live answer testing
Your question set is run against ChatGPT, Claude, Perplexity, Copilot and Google AI Overviews. Each question runs several times, because these systems are non-deterministic and a single run tells you very little. We record presence, description accuracy and every competitor cited.
- 05 Days 8 to 10.
Scoring and prioritisation
Findings are scored across the five areas and combined into an overall visibility score. Every issue is then ranked on commercial impact against implementation effort, and sorted into fix this week, fix this quarter, and fix when you next replatform.
- 06 Days 10 to 14.
Report and walkthrough
You receive the written report, then we walk through it on a call with whoever you want present. These work best with a commercial owner and a technical owner in the same room, because most findings need both to sign off.
What you actually receive
| Deliverable | What it is for |
|---|---|
| Scored visibility report | Written so a commercial director can read the first five pages and a developer can work from the rest. |
| Crawler access matrix | Every AI crawler against every relevant surface, with the response received and the rule responsible for it. |
| Structured data inventory | What exists, what validates, what is missing, per template, with example markup for the gaps. |
| Render comparison | Raw HTML against rendered HTML, showing precisely what is invisible to a non-executing crawler. |
| The AI answer log | Every question, model and run, recorded verbatim with timestamps. Raw evidence rather than a summary — clients tend to find it the most persuasive document in the pack. |
| Competitor benchmark | Named competitors scored on the same criteria, so you can see whether you are behind the sector or level with it. |
| Prioritised fix list | Ranked by impact against effort, with a plausible owner against each item. |
Working with your existing team
Most of the difficulty in this work is organisational rather than technical. Robots.txt sits with IT. The CDN sits with an infrastructure partner. Templates sit with an agency. Each group is doing its job correctly, and the gap is in the join between them.
So we work in whichever mode suits your set-up. We implement directly in your codebase or CMS. We pair with your developers and review their pull requests. Or we write specifications precise enough that your team or your agency can build from them without coming back with questions.
What we need from you
For an audit: a domain, a list of competitors, and half an hour to agree the question set. Nothing else — everything we test is publicly accessible, which is the whole point.
For implementation: repository or CMS access, and a route into whoever owns your CDN configuration. We work to your release process and your change control, including peak trading freezes.