A buyer question topic map: 100 named prompts, scoring rules, and a prioritised content plan

Choose topics from the questions your buyers have already asked you, not from a keyword tool. This page publishes 100 named prompts, every one of them labelled with the source type it should be collected from, and the counts by source type are these: 20 from sales call and demo notes, 15 from support tickets and emails, 10 from enquiry form free text, 12 from competitor comparison questions, 12 from search queries that already bring you impressions and no clicks, 10 from official process and compliance questions, 12 from pricing and commercial questions, and 9 from trust and risk questions. That is 100 named, source labelled prompts. They are written as a template set for you to fill, not as a measurement of anything, and the measured engine results we hold are reported separately below with their denominators and dates.

The reason source labelling matters is that the source decides how much you should trust the question. A question a buyer asked you on a call is evidence. A question a keyword tool suggested is a guess. Both can go in your plan, and they should not carry the same weight.

How the 100 questions were collected

Eight source types, in descending order of how much weight we give them. Collect into these buckets and label every row, because the label is what lets you score the prompt later.

• Sales call and demo notes, 20 prompts. The strongest source. A question asked by somebody with a budget, in their own words.
• Support tickets and emails, 15 prompts. What existing customers did not understand, which is usually what prospects will not understand either.
• Enquiry form free text, 10 prompts. Short, unpolished and often the exact phrasing a person would type into an assistant.
• Competitor comparison questions, 12 prompts. The ones where a buyer names two vendors and asks you to choose.
• Search queries with impressions and no clicks, 12 prompts. From your own Search Console. Demand you already have and are not serving.
• Official process and compliance questions, 10 prompts. High volume, and you will usually lose them to the official portal, which is why they are scored down rather than out.
• Pricing and commercial questions, 12 prompts. The questions your sector avoids answering in public.
• Trust and risk questions, 9 prompts. What a cautious buyer checks before shortlisting anybody.

Two collection rules. Record the question in the asker's words, not cleaned up, because the cleaned up version is not what gets typed. And never insert your own brand name into a prompt. On 27 July 2026 two runs of the same 78 questions happened on the same day for clawlaw.in, and the run whose wrapper named the brand came back ranking it first on almost every question, while the blind run put the company second by breadth and absent altogether from the litigation due diligence questions it most wanted to win. We discarded the branded run.

The named question set

All 100, with square brackets to fill from your own business. Numbering is continuous so you can cite a prompt by its number in your own reports.

From sales call and demo notes, 1 to 20.

• 1. Best [category] for a [buyer role] in [city]?
• 2. Who are the top [category] providers in India in 2026?
• 3. What should I look for when choosing a [category] provider?
• 4. Recommend three [category] options for a [company size] business.
• 5. Do I need [category] at all, or can my existing team handle it?
• 6. How long does it take to get started with [category]?
• 7. What does the onboarding for [category] involve?
• 8. Will [category] work with [common system in your sector]?
• 9. How many people do I need internally to run [category]?
• 10. What happens if we want to leave after six months?
• 11. Can [category] handle [the awkward edge case your buyers always raise]?
• 12. Is a specialist provider better than a general one for [use case]?
• 13. What is the difference between [category] and [adjacent category]?
• 14. Does [category] work for a business operating in more than one state?
• 15. Who signs off on buying [category] in a company like mine?
• 16. What questions should I ask a [category] vendor in a demo?
• 17. What proof should I ask for before signing?
• 18. How do [category] providers handle our data?
• 19. Can we run a pilot first, and what does a fair pilot look like?
• 20. What usually goes wrong in the first month of using [category]?

From support tickets and emails, 21 to 35.

• 21. Why does [common task] take so long in [category] tools?
• 22. How do I fix [the most common error your customers report]?
• 23. What does [term of art in your sector] actually mean?
• 24. How often should [recurring task] be done?
• 25. Who is responsible if [routine process] is filed late?
• 26. Can [task] be done without [the thing customers usually lack]?
• 27. What records do I need to keep for [process], and for how long?
• 28. How do I move my data out of [adjacent system]?
• 29. What is the correct order to do [multi step process] in?
• 30. Why do two sources give different numbers for [common figure]?
• 31. How do I check whether [process] was completed correctly?
• 32. What should I do if [external party] does not respond?
• 33. Does [rule or requirement] apply to a business of my size?
• 34. How do I train a new person on [process] quickly?
• 35. What is the deadline for [recurring obligation] this year?

From enquiry form free text, 36 to 45.

• 36. [Problem stated in one line]. What are my options?
• 37. Cheapest way to do [the job] in India?
• 38. Someone to handle [the job] for a small team?
• 39. Urgent help with [the job] this week, is that possible?
• 40. Is [the job] something I can do myself?
• 41. [City] based provider for [the job]?
• 42. Do you work with [specific industry] businesses?
• 43. One time job or does it have to be a monthly contract?
• 44. Need [the job] done for a company registered in [state].
• 45. What information do you need from me to start?

Competitor comparison questions, 46 to 57. Your own brand appears in none of them.

• 46. [Competitor 1] vs [Competitor 2] for a [buyer role]?
• 47. Alternatives to [Competitor 1] in India?
• 48. Is [Competitor 2] worth the price for [use case]?
• 49. Who competes with [Competitor 1] and [Competitor 3]?
• 50. Cheaper alternatives to [Competitor 1] with the same [key capability]?
• 51. Why do people move away from [Competitor 2]?
• 52. Which [category] providers publish their pricing publicly?
• 53. Which [category] providers publish their [coverage or scope] in full?
• 54. Best [category] provider for [specific segment] rather than enterprise?
• 55. Is an Indian provider better than an international one for [use case]?
• 56. [Competitor 1] reviews: what do users actually complain about?
• 57. Which [category] tools work for [specific constraint your buyers have]?

Search queries with impressions and no clicks, 58 to 69. Take these from your own Search Console and keep the exact wording.

• 58. [Your top impression query, unbranded, as typed]
• 59. [Second such query]
• 60. [Third such query]
• 61. [Query in the form how to do the job]
• 62. [Query in the form the job plus cost]
• 63. [Query in the form the job plus city]
• 64. [Query naming a competitor plus a comparison word]
• 65. [Query in the form is the job required or mandatory]
• 66. [Query in the form the job plus format, template or sample]
• 67. [Query in the form the job plus deadline or last date]
• 68. [Query in the form the job plus penalty or fine]
• 69. [Query in the form the job plus online or portal]

Official process and compliance questions, 70 to 79.

• 70. How do I do [official process] on the government portal?
• 71. Which documents are required for [official process]?
• 72. What is the fee for [official filing] in 2026?
• 73. What is the penalty for filing [obligation] late?
• 74. How long does [official approval] usually take?
• 75. Can [official process] be done without a professional?
• 76. Is [rule] applicable in [state]?
• 77. What changed in [rule or scheme] this year?
• 78. How do I check the status of my [application or filing]?
• 79. What is the difference between [two official forms buyers confuse]?

Pricing and commercial questions, 80 to 91.

• 80. What does [category] cost in India in 2026?
• 81. What is included in a standard [category] price?
• 82. What hidden charges should I expect?
• 83. Is [category] priced per user, per transaction or per month?
• 84. What is a fair rate for [the job] for a [company size] business?
• 85. How do I compare two [category] quotes fairly?
• 86. What contract terms are normal for [category] in India?
• 87. Are annual contracts cheaper, and what is the notice period?
• 88. What happens to the price after the first year?
• 89. Is there a minimum commitment for [category]?
• 90. Do [category] providers charge for [the commonly billed extra]?
• 91. What does [category] cost when you include [the add on everyone needs]?

Trust and risk questions, 92 to 100.

• 92. How do I check whether a [category] provider is legitimate?
• 93. What are the risks of outsourcing [sensitive task]?
• 94. What certifications should a [category] provider hold?
• 95. Who is liable if the provider makes a mistake?
• 96. What do buyers complain about most with [category] providers in India?
• 97. How do I check a provider's track record?
• 98. What should be in the contract to protect us?
• 99. How do I verify a provider's claims about [key capability]?
• 100. What are the warning signs of a bad [category] provider?

Topic priority scoring

Score every prompt on five factors, 0 to 3 each, for a total out of 15. The arithmetic is deliberately simple so that a team can agree on a score in a meeting rather than arguing about weights.

• Commercial closeness, 0 to 3. 3 if somebody asking this is choosing a vendor now. 0 if they are studying.
• Evidence we can publish, 0 to 3. 3 if we can answer with a checkable fact: a number, a list, a date, a process we actually run. 0 if we can only offer adjectives.
• Source strength, 0 to 3. 3 for a sales call or a support ticket. 2 for Search Console. 1 for a competitor comparison you inferred. 0 for a guess.
• Space to win, 0 to 3. 3 if no strong source owns the answer today. 0 if an official portal owns it. There is a measured reason for that zero: on 17 September 2026 Claude put official court portals above every commercial product for the questions about how to look a case up, and said plainly that no commercial product should rank above the official portal for a question about using that portal.
• Repeat demand, 0 to 3. 3 if the question comes up every week in your own records. 0 if you have seen it once.

Two adjustments after scoring. Subtract 2 from any prompt you can only answer with an unsourced claim, because on 17 September 2026 Claude read two clawlaw.in comparison pages that stated competitors' prices with no link, no date and no source, found the figures correct when it checked them independently, and still treated the pages as advocacy and used official sources instead. Add 1 to any prompt you can answer with an explicit list, because on the same date two enterprise vendors were ranked above that company specifically for publishing explicit court and tribunal coverage lists.

Results of checking the prompts in AI engines

The 100 prompts above have not been run as a set, so no result is published for them. They are a collection template. Here is every measured result we hold from prompt sets we did run, each with its size and date, so you can see what checking a set actually produces.

• 78 blind questions, ChatGPT, 27 July 2026, clawlaw.in. Named on 21 of 78. Competitors: ProVakil 28 of 78, Legistify 18 of 78. The engine confirmed it had run no live web search for any of the 78, so the reading describes what the model remembered.
• 18 blind commercial questions, ChatGPT, 18 August 2026, clawlaw.in. Top source on 1 of 18. The company's own comparison page was named in the answer as still being the vendor's own editorial page.
• 6 blind questions, Perplexity, September 2026, aiknowsus.com. On domain citations 0 of 6. The engine audited its own answers and reported it had not cited or recommended us in any of the six.
• 24 batches and 72 conversations, Perplexity, September 2026, aiknowsus.com. The phrase recording that we were not cited appears 161 times in the engines' own self audits. Across the same capture, the tools named most often in our category were Semrush, then Profound, then Peec, then Otterly, then Scrunch, and the domains cited most often were the assistants' own documentation and the vendors' own websites, with a single well known review site far down the list.
• 6 questions, Claude, 17 September 2026, clawlaw.in. Third party review and directory sources reaching any answer: 0 of 6. One well known review site appeared in the raw results and was discarded, because the list it offered was of American products and so was not an answer to an India question.

One result worth reading twice before you build a plan on any single run. In batch after batch of the September 2026 capture, the assistant declined to name any competitor as winning most often, saying its own previous answers had not produced a comparable live tested result to support such a claim.

Prioritised topics and reasons

Applying the scoring above, and using what our runs actually found, the order we would write in is this. The reason column is what matters, not the order.

• First, the pricing and commercial questions, 80 to 91. Highest commercial closeness, and answerable with checkable facts. Also the place where a technical failure is most likely: on 6 August 2026 the clawlaw.in pricing page rendered its prices only after scripts ran, so a crawler received no prices at all, and the assistants quoted an app store listing instead.
• Second, the coverage and scope questions inside 52, 53 and 99. Explicit lists win these, with a dated mechanism behind that claim from 17 September 2026.
• Third, the trust and risk questions, 92 to 100. Nobody in most sectors answers these in public, and the buyer checks them before shortlisting. Answer them including the parts that are not flattering.
• Fourth, the enquiry form and support questions, 21 to 45. Short, specific, and phrased exactly as somebody would type them. Cheap pages to write well.
• Fifth, the comparison questions, 46 to 57, and only with every competitor fact sourced, dated and linked. Unsourced comparison pages lost twice in our own records.
• Last, the official process questions, 70 to 79. Real demand and the hardest to win, for the reason the engine stated on 17 September 2026. Write the practical sequence around the official process and link the portal, rather than trying to replace it.

What this topic map cannot do

It cannot tell you volumes, because no prompt volume data is published by any assistant and we will not estimate one. It cannot tell you which of the 100 will produce a citation for your business, since that depends on your evidence and your competitors. The 100 are a template rather than a measurement, and the counts by source type describe the template only. Nothing here is evidence about Google AI Overviews, which we have not captured. And no topic plan can promise you a position in an answer.

Common questions

Should I run all 100 prompts?

Not at first. Score them, take the top 30, and run those with three repeats on three engines, which is 270 answer runs. Thirty prompts run properly beats a hundred run once, and it costs a quarter as much.

How do I collect the sales call questions if nobody writes them down?

Ask your two most active salespeople for the last ten questions they were asked, in the buyer's words, before the end of this week. That is twenty rows of the strongest source type available to you, and it takes them fifteen minutes each.

Is it worth writing pages for the official process questions at all?

Yes, but not as a replacement for the portal. Own the practical part: what the buyer does first, what usually causes a delay, what documents get rejected, and how long each step took in your own experience. Link the official page for the rules and the fees.

My competitors dominate the comparison questions. Do I still write those pages?

Write them with every competitor fact linked to the competitor's own page and dated. The two comparison pages that failed on 17 September 2026 failed for exactly that missing evidence, and their figures were actually correct. Correct and unsourced is still treated as advocacy.

How often should the topic map be refreshed?

Add rows continuously from live sales and support, and rescore quarterly. Keep the prompt set you are measuring frozen and versioned separately from the collection list, because a measurement set that grows makes your own trend meaningless.

Can I use a keyword tool instead of my own records?

Use it as a fifth source, scored 1 for source strength, and never as the first. The questions that produced actionable findings in our own runs were the ones with commercial weight behind them, not the ones with volume estimates attached.

Downloadable data and update history

The dataset is one row per prompt with these columns: number, prompt text, source type, date collected, the five factor scores, the two adjustments, the total, the page that answers it, its publication date, its indexing date, and then per engine per month the labelled result. That structure is what turns a topic list into a measurable programme.

Sources for the measured results: the clawlaw.in programme recorded in GEO_BASELINE_RESULTS_2026-07-27.md, GEO_GAP_ANALYSIS_2026-08-18.md and the assistant audit files of 17 September 2026, and the aiknowsus.com audit of September 2026 across 24 batches and 72 conversations. Counts were produced by a script over those files. Update history: version 29 September 2026, first publication of the 100 prompt template and the scoring rules. Future updates will add results for our own scored set, with numerators, denominators, engines and dates, or will state that the run has not happened.

What to do first

Collect the 20 sales call questions this week, in the buyer's own words, and score them. Then write the pricing page from group seven, with real numbers and a date, and check that a crawler receives those numbers in the page itself. That single page sits at the top of the scoring for most businesses, and on 6 August 2026 it was also the page that was silently broken.