The AI Shortlist Plays

How to use the plays.

Each play answers one question a B2B team asks about being named by AI engines, in the words the team would use. Every play page has the same parts: the effort in hours, who does it, an evidence grade with its dated source, a 10-minute self-check you can run today, numbered steps, what we saw when we did it on prolinkage.com, how to measure it, and the sources.

If you have one hour, start with the five plays in Measure first. They tell you which of the other categories to work on. If the engines describe you wrongly, go to categories 2 and 4. If they describe you correctly but name other firms, go to categories 3, 5 and 7. If you cannot tell whether any of it brings enquiries, go to category 8.

Evidence grades. Strong: published research with stated numbers, or the engine owner’s own documentation, pointing the same way. Medium: one study or primary guidance, with a reasoned link to the play. Emerging: a sensible practice that nobody has measured yet, or a vendor’s own claim. Every grade names its source and date.

Our own site. Each play says what we found when we applied it to prolinkage.com, with the date, and what we did about it. Where we have not tested it yet, the play says so. Our AI search visibility work for clients follows the same order.

Measure first.

Before you change anything, find out what the engines say now: the right question, a fixed panel, repeated runs, a score per engine and the sources they read.

1. Play 01. Which buyer question should we test first to see whether AI recommends us?
2. Play 02. How many buyer questions do we need to test AI visibility properly?
3. Play 03. Why does ChatGPT give a different answer every time we ask?
4. Play 04. How do we score our visibility across ChatGPT, Perplexity and Google's AI?
5. Play 05. Which sources do AI engines cite when they recommend suppliers in our category?

Facts on your own site.

Give the engines one clear, consistent record of who you are, what you sell, for whom, how you price and why you can be trusted.

6. Play 06. Do we need a company facts page for AI engines to read?
7. Play 07. Do our service pages say plainly who we are, what we sell and who it is for?
8. Play 08. Should we publish our pricing approach so AI can describe it?
9. Play 09. How do we show our team and credentials so AI engines can verify them?
10. Play 10. Are our company name, founding year and service names the same everywhere?

Pages AI searches retrieve.

The engines turn a buyer’s question into several shorter searches. These plays make sure your pages answer the ones that matter.

11. Play 11. Do we have a page that answers the 'best X for Y' searches AI runs for our category?
12. Play 12. Should we publish comparison and alternatives pages?
13. Play 13. Which use-case pages do AI engines retrieve when buyers describe their problem?
14. Play 14. Do FAQ blocks still help our pages get quoted by AI?
15. Play 15. How do we write sections that AI engines can quote as they stand?

Profiles the engines read.

Directories, LinkedIn, Google Business Profile and partner listings often describe you before your own site does. Make them agree.

16. Play 16. Which review directories do AI engines read for our category?
17. Play 17. Does our LinkedIn company page change what AI says about us?
18. Play 18. Does a B2B firm need a Google Business Profile for AI search?
19. Play 19. How should we use partner listings and badges such as Google Partner?
20. Play 20. Do our facts match across every profile the engines read?

Proof engines can quote.

Results only count if they can be checked: dated case studies, a public method, numbers with units, and client names used with permission.

21. Play 21. How do we write case studies that AI engines can quote?
22. Play 22. Should we publish how we work as a public method page?
23. Play 23. How do we make our results quotable, with units and dates?
24. Play 24. Can we name our clients in public proof, and how do we ask?
25. Play 25. How do we stop AI engines quoting our old, stale numbers?

Access and structure.

If a crawler cannot reach or read a page, nothing else in this guide matters. These plays check access, rendering, markup and layout.

26. Play 26. Can AI crawlers actually reach our website?
27. Play 27. Can AI engines read our content without running JavaScript?
28. Play 28. Does our schema match what the page actually says?
29. Play 29. Do our headings tell AI engines what each section answers?
30. Play 30. Are pop-ups or cookie walls blocking AI engines from our pages?

Lists and third-party mentions.

Most of what engines repeat about a company comes from other people’s pages. Earn a place on them honestly, never by paying for disguised editorial.

31. Play 31. How do we get into 'best agencies' roundup lists honestly?
32. Play 32. Which trade press mentions do AI engines pick up?
33. Play 33. Should we take part in Reddit and community threads about our category?
34. Play 34. Do expert quotes and guest contributions help us get named by AI?
35. Play 35. Which third-party pages mention our competitors but not us?

From AI exposure to pipeline.

Visibility is not the goal. These plays connect AI referrals and mentions to enquiries, pipeline and closed revenue, and report each separately.

36. Play 36. How do we track AI referrals in GA4?
37. Play 37. Should we ask leads how they heard about us and store it in the CRM?
38. Play 38. How do we measure branded search and direct lift from AI exposure?
39. Play 39. How should we report AI visibility, behaviour and pipeline separately?
40. Play 40. How do we connect AI-referred leads to closed revenue?

What the plays are not.

They are not a promise. Nobody outside the engines controls what ChatGPT, Perplexity or Google’s AI says, and answers change from one run to the next. The plays remove the reasons an engine could not name you, and give you a way to measure whether that is working.

They are not a shortcut. None of the plays involves paying for a place on a list presented as editorial, writing reviews for clients, or publishing a ranking that puts yourself first. Those tactics are easy to spot, and they put the rest of your evidence in doubt.

They are dated. Engines, studies and our own results change. Each edition is dated, each play shows when it was written and what it cites, and the changelog above records what changed between editions.

Sources for the evidence grades are listed on each play page. Edition 1 cites research published between August 2023 and July 2026 and our own observations from 6 to 8 November 2026.