agentleFS
frontmost-aifrontmost-ai/aeo-skills/skills/customer-survey/SKILL.md

customer-survey

Skill1 starsChanged 3 days agoMITView on GitHub

Find out whether AI apps send a business customers. Adds one optional question, "Where did you first hear about us?", with AI assistants as their own answer, and reads the answers beside analytics. Includes a ready prompt for a coding agent. Use for "is AI sending me customers", "how do I measure ChatGPT traffic", or AI attribution. Needs no Frontmost AI account.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 12 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈12

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask "Where did you first hear about us?" once, at the first moment that fits:

    line 14
  2. Distinctive:

    Use ten answers or fewer, one for each row.

    line 26
  3. Distinctive:

    When a customer picks "An AI assistant", ask "Which one?": ChatGPT, Claude, Gemini, Google AI Mode or AI Overviews, Perplexity, Microsoft …

    line 28
  4. Distinctive:

    Use the channels that this business's buyers use.

    line 29
  5. Distinctive:

    Keep the answer where the payment status is, so that paying customers can be counted by source.

    line 38
  6. Distinctive:

    In the product's code: If you work in the code and the person asks, follow build-prompt.md.

    line 42
  7. Distinctive:

    Read the answers after a few weeks.

    line 48
  8. Distinctive:

    Read three signals together.

    line 50
  9. Distinctive:

    When mentions rise and more paying customers pick an AI app, both signals moved in the same direction.

    line 60
  10. Very specific:

    When mentions rise and the survey answers do not move, read what the AI apps say about the business.

    line 61
  11. Distinctive:

    Do not make the question required.

    line 73
  12. Distinctive:

    Do not ask it again after an answer or a skip.

    line 74
  13. Distinctive:

    Do not collect more than the answer, the AI app, and the "Other" text.

    line 75

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Customer survey

A count of AI mentions shows whether AI apps mention a business. It does not show whether those mentions bring customers. Analytics miss much of it: a buyer hears of the business in ChatGPT, then searches for its name, and the visit looks like a search visit.

So ask the customers. Use this guidance to add one question, to save the answers correctly, and to read them.

1. Place the question

Ask "Where did you first hear about us?" once, at the first moment that fits:

BusinessWhere to ask
Software with sign-upThe first screen after sign-up
Online storeDirectly after the first purchase
Service businessThe enquiry form, or the first call. The person who takes the call records the answer

Make the question optional. It takes one tap. A customer sees it once.

2. Write the answers

  • Use ten answers or fewer, one for each row.
  • Make "An AI assistant" one of the answers. Without it, a buyer who found the business in ChatGPT picks "Google" or "Other".
  • When a customer picks "An AI assistant", ask "Which one?": ChatGPT, Claude, Gemini, Google AI Mode or AI Overviews, Perplexity, Microsoft Copilot, Grok, or another.
  • Use the channels that this business's buyers use. A store can replace LinkedIn with TikTok or Instagram. A local business can add Google Maps.
  • Include "Other" with a short text field.

A starting list: An AI assistant. Google or another search engine. YouTube. LinkedIn. Reddit. X (Twitter). Other social media. A friend or colleague. Podcast, newsletter, or article. Other.

3. Save the answers

  • Save each answer on the customer's record, with the date.
  • Save a skipped question as skipped, with no answer. Never save a skip as "Other".
  • Keep the answer where the payment status is, so that paying customers can be counted by source.

4. Build it

  • In the product's code. If you work in the code and the person asks, follow build-prompt.md. List the files and the fields first, and wait for approval before you change the code.
  • In a form tool or a store platform. Give the person the question, the answers, and the rule for skips. The person adds them in that tool.
  • By hand. For a service business, add a column to the customer list, with the same answers.

5. Read the answers

Read the answers after a few weeks. Report counts, not shares alone: "7 of 52 paying customers picked an AI assistant: ChatGPT 5, Perplexity 2. 9 skipped."

Read three signals together. Each one answers a different question.

SignalQuestion that it answersWhere it comes from
The surveyWhere did the customer hear of the business?The saved answers
AnalyticsWhere did the visit come from?Visits whose referrer or utm_source is an AI app, such as chatgpt.com or perplexity.ai
AI mentionsHow often do AI apps mention the business?A mention-check count, or a Frontmost AI project
  • Count one customer once. A customer who picks ChatGPT and arrived from a search for the business name is one customer. Keep both facts.
  • Count paying customers apart from sign-ups.
  • When mentions rise and more paying customers pick an AI app, both signals moved in the same direction. This does not prove a cause.
  • When mentions rise and the survey answers do not move, read what the AI apps say about the business. Use the mention-check skill.

Limits

Say these in the result:

  • People forget, and they pick the closest answer. In a 2024 study by HockeyStack of answers from B2B buyers, 2,646 of 4,939 "search" answers matched the last visit that the software recorded. See the study.
  • The answers alone do not prove what caused a purchase. They are one signal of three.
  • A small number of answers shows little. Give the count each time.

Rules

  • Do not make the question required.
  • Do not ask it again after an answer or a skip.
  • Do not collect more than the answer, the AI app, and the "Other" text.

Output

  1. Where the question goes, and the list of answers for this business.
  2. The fields to save.
  3. The build path: the plan for the code, or the text for the form tool.
  4. The date of the first reading, and the three signals to compare.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.