AIVisCity Weekly #27: Build an AI Visibility Score You Can Actually Use
Welcome to AIVisCity Weekly
A free weekly briefing helping small business owners understand how AI search is changing customer discovery, and what practical steps they can take next.
In this week’s issue:
What’s Happening in AI Search — conflicting public information can distort AI answers, while AI shopping continues to introduce paid recommendation formats.
Weekly Insight — build a DIY AI Visibility Score around the customer questions that matter to your business.
Try This Yourself — score one genuine customer question in two AI tools to create your first mini-baseline.
Worth Reading — Google’s official guidance on generative Search and a practical guide to connecting visibility with business outcomes.
AIVisCity Answers — why a small business should track one or two relevant AI platforms separately.
🔍 What’s Happening in AI Search
Conflicting public information can produce confident but outdated AI answers
In a recent Search Engine Journal analysis, consultant Carolyn Shelby argues that the risk is often not too little brand information but too many competing versions of it. AI search may retrieve and combine sources into one answer, so an old PDF, former service description or stale profile can win when its wording more closely matches the customer’s question. For a small business, this reinforces why a visibility check must score accuracy as well as presence. Review the facts across your website, profiles and older public material; a mention that sends a customer to the wrong service is not progress.
Source: Search Engine Journal
AI shopping is beginning to mix organic and paid recommendations
A TechRadar opinion piece argues that paid placements are moving closer to the recommendation moment in AI shopping. It points to Amazon’s own reported results for Sponsored Prompts in Alexa for Shopping, while acknowledging that these figures are not an independent benchmark. The immediate lesson for an online seller is to separate two questions: are you being organically surfaced as a suitable option, and are you paying to appear in a commercial placement? As AI shopping develops, those are likely to need different budgets and measurements.
Source: TechRadar
🔮 Weekly Insight: Build an AI Visibility Score You Can Actually Use
AI Visibility is attracting service providers of dashboards, metrics and “share of voice” figures. In this week's insight, we would like to propose a simple "DIY" AI Visibility Score that every small businesses can build without expensive tools or technical knowledge.
Start with a score based on customer questions that could lead to an enquiry. It will not measure every AI conversation. A useful AI Visibility Score measures the questions your customers actually ask. It gives you a repeatable baseline for discovery, accuracy and a useful next step.
1. Choose ten customer questions
Write ten questions a customer might ask before they know your business name. Use their language and cover the work you want to win.
Use this 4–3–2–1 mix:
4 discovery questions — “Best accountant for freelancers in Bristol”
3 problem questions — “Who can help a designer with a first self-assessment tax return?”
2 comparison questions — “Should a new limited company use an accountant or accounting software?”
1 trust question — “Which Bristol accountants understand limited companies?”
These questions test different customer situations, rather than merely whether AI knows the firm exists.
Save the exact wording in a spreadsheet. Revise only prompts that do not reflect a real customer; a fixed set makes the next result comparable with the first.
2. Run the same check in two AI tools
Ask each question in the same two accessible AI tools and record the date, the answer and any sources or links shown. You will have 20 answers: ten questions across two tools. Answers can vary by tool and over time, so treat one run as a snapshot, then repeat the same check at a regular interval (e.g. monthly).
3. Award up to five points for each answer
Use the same rules for every answer:
Mention — 0 points: you are absent. 1 point: you are a relevant option. 2 points: you are clearly recommended or the answer explains why you fit.
Accuracy — 0 points: you are absent or a material fact is wrong. 1 point: the description is broadly correct. 2 points: the important service, location or specialism details are correct.
Next Step — 0 points: no useful route forward. 1 point: the correct website, contact route or reason to choose you.
An absent answer scores zero overall, and a wrong location or service earns zero for Accuracy. Being visible in a misleading answer is not a win.
Twenty answers at five points each create a score out of 100. This is your DIY Simplified AI Visivility Score.
A commercial platform may look more scientific because it draws on far more prompts and a proprietary calculation, but that does not automatically make it more useful to you. This score shows exactly what went into it, keeps the customer questions in view and gives you a baseline you can repeat.
Next week, we will look at how to use these scores to improve your AI Visibility over time.
💻 Try This Yourself
Choose one question a new customer might ask without knowing your business name. For example: “Who can help with [problem] in [location]?” Ask it in two AI tools, then copy each answer into a simple note or spreadsheet. Give each answer up to two points for Mention, two for Accuracy and one for a Next Step. Your result is a ten-point mini-score. Keep the exact question and date beside it.
You have not measured your whole AI Visibility yet, but you have created a repeatable check that can become the first line of your scorecard.
📕Worth Reading
If you're curious to look deeper into how AI search is changing the internet, these articles are worth a look.
Google advises building on useful content and Search fundamentals
Google’s official guide says its generative Search features rely on core Search systems. It recommends crawlable, helpful and original content, sound technical foundations and accurate local-business or product details, while warning against special AI files and inauthentic mentions. Use it as a filter when deciding what to improve after your first score.
Read: Google’s guide
Visibility figures need a route to business outcomes
Semrush’s product-led guide makes a useful distinction: mentions and citations are early visibility signals, while leads, sales and revenue need separate evidence. It recommends a fixed set of buyer prompts, then monitoring demand and conversions over time. Use that sequence to avoid treating a score as the commercial result.
Read: the Semrush guide
✅ AIVisCity Answers
Q: Should I Track AI Visibility on Different Platforms Separately?
A: You should track platforms separately, but you do not need to track every AI tool. Start with one or two places your customers are most likely to use, plus Google AI Overviews if Google is important for your business. A simple check is more useful than one broad score.
Want a more detailed explanation? Check out here.
👋 Until Next Week
If you find AIVisCity useful and want an easier way to keep up with practical AI Visibility updates for small businesses, you can use the link below to add AIVisCity as a Preferred Source in Google Search.
How are you seeing AI affect the way people search for businesses in your industry?
If you have noticed changes — or if you tried the quick check in this issue — we’d love to hear your observations. Feel free to share them in the comments.
See you in the next issue of AIVisCity Weekly.
Comments