The service
What I do
I find the question a business can own in AI answers, then make it the answer. First its information becomes easy for assistants to find, trust and quote. The work runs as a 90-day sprint, then a monthly retainer, because what assistants say about you keeps changing even when your site doesn't.
What are the four strands of work?
01
Technical foundation Prerequisite
- AI crawler access in
robots.txtand at your CDN or firewall, often blocked by default - Key information moved out of images, JavaScript and PDFs into plain HTML, usually the most valuable single fix
- Structured data, only where it matches the visible page
- Sitemaps submitted to Google Search Console and Bing Webmaster Tools
02
The question you can own The real work
- Find the question where you can be named first: the thing you do that competitors don't, that buyers with budget actually ask
- A map of your buyers' real questions, from support tickets, sales calls, site search and reviews, ranked by how ownable each one is
- Priority pages that each answer one question, opening with a direct, quotable answer
- Your vocabulary aligned with your buyers', where the two differ
03
Entity & consistency
- One accurate description of your company across your site, directories, listings and profiles
- Contradictions found and corrected: models trust a business described one way, not five
04
Measurement
- A fixed set of buyer prompts, agreed at the start
- Run by hand in ChatGPT, Perplexity and Gemini monthly, screenshotted and dated
- Same method as the free check, so every report compares to the same baseline
Not every business has an obvious angle. When the check doesn't find one, it says so, and the work stays on the foundations and the broad questions.
What are the steps from first contact?
- 01
Free check
About an hour of hands-on research, delivered within three working days. A dated one-page report with screenshots, and the one fix to make first. What it covers
- 02
90-day sprint
Paid, fixed fee. Foundation, answer content and consistency, with baseline measurement, over three months.
- 03
Monthly retainer
Ongoing testing, reporting and new or refreshed pages, because assistants keep changing their answers.
How does an engagement run?
Month 1
Foundation
- Technical audit
- Crawler access
- Facts into HTML
- Structured data
- Baseline set
Months 2–3
Build
- Find the ownable question
- Answer pages built out from it
- Pages restructured
Months 4–6
Expand & measure
- Monthly citation tests
- Narrow → broader terms
- Competitor watch
- Rework what didn't move
Month 7 →
Retainer
- Monthly report
- New & refreshed pages
- Technical monitoring
| When | Phase | What happens |
|---|---|---|
| Month 1 | Foundation | Technical audit and fixes, key content into HTML, structured data, consistency clean-up, baseline measurement |
| Months 2–3 | Build | Buyer question map, new answer pages, existing high-traffic pages restructured |
| Months 4–6 | Expand and measure | More pages, prioritised by what actually got cited; monthly citation testing; pages that didn't move reworked, with an honest account of why |
| Month 7 on | Retainer | Monthly testing and reporting, new and refreshed pages, technical monitoring, quarterly review |
There are review points at the end of month 1 and month 3, and either of us can stop there.
What do I guarantee, and what don't I?
Guaranteed: every deliverable ships on the date we agreed, and every monthly report is accurate, including the months nothing moved.
Measured and reported, not promised: citation rate, answer-box ownership, rankings and traffic. Google and the model providers control those, not me. Anyone who guarantees you a percentage on them is guaranteeing something they don't control.
What isn't included?
- Paid media
- Link building, guest-post placement or press-release distribution
- Review seeding or planted mentions on forums, Reddit or review platforms. Those platforms penalise the brand, not the consultant, and it isn't work I'll put my name to
How do I work?
Research and drafting are AI-assisted. Every factual claim is checked by hand against the raw source, and citation results come from actually running the prompts, not from a tool that estimates them. Fixes are delivered either directly in your CMS or as a precise spec your developer can implement.
Not sure it's worth it? The free check will tell you where you stand before you spend anything.