The Scoreboard What the engines actually returned
We asked two live AI engines the questions your buyers ask, then counted how many times you were named. The headline number is stark, but read it carefully, because the shape of it is the opportunity.
0 of 11 category queries cited Mads. The only hit was the branded query "Mads Singers" - your own name.
Same result, independently. 0 of 11 category queries, 1 branded hit. Two engines, one answer.
Across every "I need a coach" style question, on both engines, you were surfaced zero times.
The plain-English version: the only query that returns Mads is the one where someone already types "Mads Singers". For all 11 of the questions a stranger would actually ask, neither engine mentioned him. The 1 out of 12 is just his own name reflected back.
Real data. 12 queries x 2 engines, both with live web search, run through the DataForSEO LLM Responses API on 2026-06-12. The full query-by-query receipts are on 06 - Live AI Test.
The Key Insight The AI knows him. It just does not recommend him.
This is the single most important thing on the whole orb, so it gets its own card. The zero score is not because the AI has never heard of Mads. It is the opposite.
When asked about him by name, ChatGPT returns a clean, accurate, on-brand description, unprompted:
"Mads Singers is a management coach and consultant known for his expertise in effective management, particularly in the areas of productivity and business leadership. He often shares insights on how to optimize team performance and improve managerial effectiveness."
The entity exists and it is accurate
The model has formed a correct mental picture of who Mads is and what he does. That is the expensive half. Most experts never even get this far. The raw recognition is there.
Recommendation eligibility
Knowing who someone is and recommending them are two different jobs for an AI. Mads passes the first and never reaches the second. That bridge is what is missing, not the knowledge.
This is a surfacing gap, not a knowledge gap, and that is the best possible kind to have. A knowledge gap means the AI has never encountered you and you start from zero. A surfacing gap means the recognition is already built; you just have not been positioned as a thing to recommend. It is the most fixable problem in this entire report.
Why It Matters Every category answer is a buyer you did not meet
These are not abstract test queries. They are the literal questions your ideal clients type, often word for word, when they are ready to find help. Here is who the engines hand them to instead.
What your buyer asks
- best people management coach
- management coach for entrepreneurs
- how to delegate effectively as a business owner
- executive coaching for delegation
- DISC assessment training expert
Who the AI hands them to instead
- Marshall Goldsmith, Simon Sinek, John C. Maxwell
- Brene Brown, Liz Wiseman
- Thomas International, Everything DISC, Discflow
- MentorCruise, Tony Robbins, FutureLearn
Think about who is in that right hand column. They are not better at the work than Mads. Some of them are authors who do not coach one to one at all. They win the answer because they are positioned to be recommended: a consistent named entity, structured profiles the engine can lean on, and a steady stream of cited content. That is a machine-readability advantage, not a coaching advantage.
The cost is quiet and continuous. Every time one of those 11 questions gets asked, somewhere a person who would have been a great fit for EPM Mastery gets sent to a book or a platform instead. You never see it happen. There is no bounced email, no missed call. The buyer simply meets someone else first.
The engines built those category answers mostly off YouTube (it appeared roughly seven times across the source lists) plus a handful of UK coaching and training sites. Mads has a real YouTube presence and a content-heavy site, so the raw ingredients to be one of those sources already exist. They are just not yet structured or positioned to be picked up as a recommendation source. The supply is there; the packaging is the work.
The Fix Direction How a surfacing gap gets closed
Because this is a surfacing problem and not a recognition problem, the fix is targeted rather than enormous. It comes down to three layers, all of which build on the entity that already exists.
Entity
Make the machine-readable "who is Mads" signal consistent and site-wide, with Person and Organization schema and clean sameAs links to the YouTube, LinkedIn and podcast presence. Turn the accurate description into a structured fact the engines can cite with confidence.
Answer content
Build content that directly answers the 11 buyer questions in a citable, well-structured way, so that when the engine assembles its answer, a Mads Singers page is the most useful thing to reach for. You already have the expertise; this gives it the shape an AI will quote.
Platform presence
Strengthen the third-party signals (profiles, directories, the YouTube footprint the engines already trust) so the recommendation has corroboration. This is what tips an entity from "known" to "worth suggesting".
None of this is a rebuild. It is a re-positioning of assets that mostly already exist, aimed squarely at the gap the data just exposed. The full, sequenced version of these three layers lives on 12 - GEO Playbook, with the 90 day order of operations on 13 - 90-Day Map.
A note from a friend, not a pitch. We ran this because Mads coaches Donal on the business, and turning the agency's machine on his own site felt like a fair thank you. The finding is genuinely good news dressed as a zero: the hardest part, getting the AI to understand you, is already done. If any of this is useful, brilliant. If not, no pressure at all. It is yours to keep either way.