AI Visibility - Is your business invisible to AI search?
When a buyer asks ChatGPT "who is the best renovation specialist in Dundee" or Perplexity "find me an AI automation agency in Scotland," the answer comes from somewhere. A business that has done the right groundwork gets named. One that has not, no matter how good its Google ranking, is simply absent.
01
That gap is what AI visibility work closes. Assist IQ audits where your business stands today, maps what it would take to be consistently named, and manages the inputs the evidence says actually move the needle.
02
What is AI visibility, AEO and GEO?
AEO (answer-engine optimisation) is appearing in the answers AI assistants give buyers. GEO (generative-engine optimisation) is structuring content so generative engines retrieve and cite it. In practice the two overlap: get the fundamentals right and both improve.
ChatGPT, Perplexity, Google AI Overviews and Copilot all pull from the same places: Bing's index, review platforms, directories, third-party editorial mentions, and your own website copy. What changes the outcome is being present, parseable and corroborated in those places.
Classic SEO gets you eligible. Third-party corroboration gets you chosen.
[Assist IQ measures its own AI visibility using the same method we sell. Baseline run July 2026.]
03
Why it matters now
Buyers now use answer engines to shortlist suppliers, compare approaches and frame the questions they bring to a sales call. The channel is still changing, which is why Assist IQ treats it as a measured visibility discipline rather than a guaranteed source of traffic.
04
What actually moves the needle
Most agencies selling AI visibility focus on easy-to-charge-for work: schema markup, llms.txt files, technical audits. The evidence is more specific.
The two levers we prioritise:
- Review signals. Recent, credible reviews help a buyer and an answer engine corroborate what a business claims about itself.
- Third-party mentions. Directories, editorial comparisons, trade bodies and local press give an engine sources beyond the business's own website.
What we do not sell as primary levers:
- llms.txt files. We may ship one as a low-cost courtesy to machines. We never claim that it drives citations.
- Schema as a ranking mechanism. It helps machines read facts; Google is explicit that it does not affect AI selection.
- Guaranteed AI rankings. Answers vary between engines and runs. We report repeated mention patterns, not a single favourable answer.
The agency that tells you what does not work is the one worth trusting with what does.
05
The method: a /100 scorecard and six pillars
Every AI Visibility engagement starts with the same structured audit. Six pillars, one clear insight each, combined into a weighted score out of 100.
| Pillar | What we assess | Weight |
|---|---|---|
| A. Answer-surface baseline | Run your key queries across ChatGPT, Perplexity, Google AI Overviews and Copilot. Who gets named? What sources get cited? Is anything factually wrong about your business? | 15 pts |
| B. Citation-source coverage | Which domains did the engines retrieve when answering those queries? Are you present in those sources? The gap between "where engines look" and "where you appear" is the most actionable finding in the audit. | 20 pts |
| C. Review and proof signals | Count, recency, platform spread and response rate versus your competitors. Proof trapped somewhere engines cannot see is the most common problem we find. | 25 pts |
| D. Entity and machine-readability | Google Business Profile completeness, Bing indexation, AI crawler access, NAP consistency across all directories, basic schema. | 20 pts |
| E. On-site answerability | Does your website answer the questions your buyers are putting to AI assistants? Answer-first structure, comparison content, named statistics, FAQ blocks. | 10 pts |
| F. Competitive reverse-engineering | For every competitor the engine names over you: why them? That becomes the plan. | 10 pts |
Score bands:
- 0 to 30: Invisible band. Absent from AI answers for almost all relevant queries.
- 31 to 55: Eligible but unchosen. The engines can find you; they are choosing others.
- 56 to 75: Contender. Being named in some contexts, with clear levers left to pull.
- 76 and above: Incumbent. Strong citation presence; focus shifts to defending and widening.
The scoring weights follow the evidence: review signals and citation coverage together account for 45 points. The technical pillar that most agencies lead with accounts for 10.
06
What is included
| Component | What it covers |
|---|---|
| AI Visibility Blueprint (one-off) | Full six-pillar audit, weighted scorecard, before-state baseline captured and dated, written client report, 5 to 10 page on-site GEO build-spec (answer blocks, FAQ structure, comparison content, schema), citation-coverage gap list in evidence-ranked order, GBP and Bing optimisation checklist |
| Monthly monitoring | Pillar A battery re-run across three AI engines (three samples each, mention rates reported), "what moved" summary, review-count and coverage-ratio delta, quarterly full re-audit |
| Review engine | Automated post-job review request via the CRM, owner-response cadence, fragmented-proof consolidation |
| Citation build-out | Directory listings in vertical-staple order (the sources engines actually retrieve), local press pitch drafts, NAP consistency sweep |
The Blueprint is a one-off engagement. Monitoring and the review engine run inside the ongoing retainer.
The £750 AI Ops Review can include the initial visibility diagnosis where it is part of the operating problem. Any blueprint, implementation or monitoring work is then scoped from the evidence. We do not promise a ranking or sell an unsupported score improvement.
07
Proof: we measure ourselves
Assist IQ uses the same method on itself. We run the same six-pillar audit on our own business, against the same queries our prospects are asking AI assistants. We track our own mention rates across ChatGPT, Perplexity and Google AI Overviews on a monthly basis.
We do this for two reasons. The first is that it keeps us honest: if we cannot move our own needle, we have no business charging clients to move theirs. The second is that it demonstrates the method in the most credible possible way. We are not selling a theory.
ScottishWill is an operational reference for Assist IQ's wider work, but we do not publish an AI-visibility result for it without a dated, repeated baseline and follow-up run. The method is designed to make that distinction visible.
08
Honest limits
We cannot guarantee that an AI names your business by a specific date. Nobody can. We guarantee the inputs the evidence says move the odds, and a measured before/after that shows what changed.
If you need results next week from the biggest traffic channel possible, paid search is a faster lever and we will tell you that. If you are building for where the highest-intent buyers are going over the next 12 to 24 months, this is the right work to start now.
The highest-value actions in most audits are mechanical: reviews up, directories filled in, site copy that answers the right questions. Less impressive to talk about than proprietary AI ranking algorithms. More effective in practice.
09
Questions we get asked
How is AI visibility different from SEO?
SEO focuses on visibility in conventional search results. AI visibility focuses on whether answer engines can understand, retrieve and cite a business in a generated response. The foundations overlap: crawlable pages, clear entities, useful answers and credible third-party evidence. AI visibility adds repeated answer-engine testing and citation analysis. Neither discipline can guarantee a ranking or mention.
Do I need a new website to improve AI visibility?
Not always. The on-site GEO work (answer-first copy, FAQ blocks, comparison content, schema) can be added to an existing site. What does matter is that your site content is rendered in HTML that AI crawlers can read without JavaScript. Our audit checks that in the first pass.
If you are in the process of building or rebuilding a site, that is actually the best time to do this work. Baking GEO in from the start costs nothing extra and makes every page a better candidate for citation from day one.
How long before we start seeing results?
The review engine and GBP work start compounding from week one. Citation-source coverage builds over months one to three. The Pillar A battery is the honest measure: we run it at baseline and again at 90 days, and the before/after is the deliverable.
The narrow-query approach gives the work a measurable starting point without promising a result by a particular date. We find a query framing the client can credibly own, record the baseline and widen only when repeated runs support the move.
Is this worth it for a small local business?
For a local business with real review material and a service people genuinely need, yes, it tends to be the highest-leverage marketing spend available. The competition for AI citation in most UK local markets is close to zero. A plasterer in Angus, a letting agent in Aberdeen, a salon in Dundee: almost none of them have done this work. The first mover in a local vertical owns that position.
The caveat: you need real proof to point to. A business with no reviews and no satisfied customers is not ready for AI visibility work. Build the proof first, then build the citation infrastructure around it.
10
Start with a baseline
The first step is understanding where you stand. An AI Visibility Blueprint gives you the scored baseline, the gap list, and the prioritised plan.
If you want to see the method working before you engage, book a AI Ops Review. We will run a quick version of the Pillar A and B checks on your business before the call and show you what we find.
A real, evidence-led look at where you stand.
11
Related services: AI Employees · Managed Operations · How AI automation agencies compare · Proof: what we have built and measured
Assist IQ is a trading name of PCMG Consulting Services Ltd, registered in Scotland (SC852814). Based in Arbroath, Angus. We serve businesses across Scotland and the UK.
Tell us about your business
Tell us what keeps landing back on your desk. We’ll look at how the work really moves, where it gets stuck, and which part is worth fixing first.
The AI Ops Review is £750. It gives you a written plan, a clear order of attack, and an honest view of what should stay human.
No prep. No data handover. Start with the problem, not the software.