AI answer audit prompt
An AI answer audit prompt reads answers from ChatGPT, Perplexity and Google's AI to your buyers' questions and records whether you're named, how you're described, who else is named, and why. It shows what to fix rather than a rank.
- Stage
- Grow, step 35 of 40
- Use it
- Once before you change anything, then monthly for your ten most important buyer questions.
- You need
- Ten buyer questions, written with a situation and a place
- Answers copied from each tool, three runs per question
- You get
- A table per answer and a list of gaps to fix on your pages. Output schema included.
The prompt
I'll paste answers from ChatGPT, Perplexity and Google AI Mode
to the same buyer question: "[question]".
For each answer, record whether [business] is mentioned, how it's described,
which competitors are named, the reason given for each,
which sources are cited, and which factor decided the answer:
use case, selection criteria, audience, credibility or brand.
Then list what my pages would need to say for me to be named,
using only gaps you can see in these answers.
[paste answers]JSON version, for teams and automations
Fill each value under "inputs", keep the rest. It follows the prompt schema. When to use JSON.
{
"$schema": "https://websitesdesigner.com/insights/prompts/prompt.schema.json",
"id": "ai-answer-audit",
"name": "AI answer audit prompt",
"stage": "grow",
"step": 35,
"of": 40,
"version": "2026-11-19",
"url": "https://websitesdesigner.com/insights/prompts/ai-answer-audit/",
"inputs": {
"question": "[question]",
"business": "[business]",
"paste_answers": "[paste answers]"
},
"instructions": [
"I'll paste answers from ChatGPT, Perplexity and Google AI Mode to the same buyer question: \"{{question}}\".",
"For each answer, record whether {{business}} is mentioned, how it's described, which competitors are named, the reason given for each, which sources are cited, and which factor decided the answer: use case, selection criteria, audience, credibility or brand.",
"Then list what my pages would need to say for me to be named, using only gaps you can see in these answers.",
"{{paste_answers}}"
],
"output": "A table per answer and a list of gaps to fix on your pages.",
"checks": [
"Every competitor mention has a reason recorded.",
"Sources cited are listed with URLs.",
"Gaps point to specific pages."
],
"output_schema": {
"type": "object",
"properties": {
"question": {
"type": "string"
},
"answers": {
"type": "array",
"items": {
"type": "object",
"properties": {
"tool": {
"type": "string"
},
"run": {
"type": "integer"
},
"mentioned": {
"type": "boolean"
},
"description": {
"type": [
"string",
"null"
]
},
"competitors": {
"type": "array",
"items": {
"type": "object",
"properties": {
"name": {
"type": "string"
},
"reason": {
"type": "string"
}
},
"required": [
"name",
"reason"
],
"additionalProperties": false
}
},
"sources": {
"type": "array",
"items": {
"type": "string"
}
},
"deciding_factor": {
"type": "string",
"enum": [
"use case",
"selection criteria",
"audience",
"credibility",
"brand association"
]
}
},
"required": [
"tool",
"run",
"mentioned",
"description",
"competitors",
"sources",
"deciding_factor"
],
"additionalProperties": false
}
},
"gaps": {
"type": "array",
"items": {
"type": "object",
"properties": {
"gap": {
"type": "string"
},
"page": {
"type": [
"string",
"null"
]
}
},
"required": [
"gap",
"page"
],
"additionalProperties": false
}
}
},
"required": [
"question",
"answers",
"gaps"
],
"additionalProperties": false
}
}Output schema (JSON Schema)
Ready for strict structured-output modes: every field required, unknown values nullable. In a chat, paste it after the prompt with "Return only JSON that matches this schema." How output schemas work.
{
"type": "object",
"properties": {
"question": {
"type": "string"
},
"answers": {
"type": "array",
"items": {
"type": "object",
"properties": {
"tool": {
"type": "string"
},
"run": {
"type": "integer"
},
"mentioned": {
"type": "boolean"
},
"description": {
"type": [
"string",
"null"
]
},
"competitors": {
"type": "array",
"items": {
"type": "object",
"properties": {
"name": {
"type": "string"
},
"reason": {
"type": "string"
}
},
"required": [
"name",
"reason"
],
"additionalProperties": false
}
},
"sources": {
"type": "array",
"items": {
"type": "string"
}
},
"deciding_factor": {
"type": "string",
"enum": [
"use case",
"selection criteria",
"audience",
"credibility",
"brand association"
]
}
},
"required": [
"tool",
"run",
"mentioned",
"description",
"competitors",
"sources",
"deciding_factor"
],
"additionalProperties": false
}
},
"gaps": {
"type": "array",
"items": {
"type": "object",
"properties": {
"gap": {
"type": "string"
},
"page": {
"type": [
"string",
"null"
]
}
},
"required": [
"gap",
"page"
],
"additionalProperties": false
}
}
},
"required": [
"question",
"answers",
"gaps"
],
"additionalProperties": false
}What good output looks like
- Every competitor mention has a reason recorded.
- Sources cited are listed with URLs.
- Gaps point to specific pages.
What experts miss
- Run each question three times per tool. Answers vary between runs, so one screenshot is an anecdote.
- Track the reasons, not only the names. The reason an assistant gives for a competitor is the content you're missing.
- Ask with the buyer's setup in the question. "For a dental clinic on Wix in Tallinn" gets a different answer than "best web designer", and it's closer to what buyers type.
Why it works
AI answers drift as models and sources change. Recording why each competitor was named shows the content you're missing. A rank tells you only that you lost.
How to adapt it
- Add the sources each tool cited to an outreach list.
- Run the same audit for a competitor to see what they're credited for.
Mistakes to avoid
- Asking with your brand name in the question.
- Changing pages after one run.
Go deeper
See AI answer monitoring and why AI citations are the new impressions. Plan fixes with the determinants map. This prompt is part of the grow stage. It follows the AI website design prompts guide and the six steps in how to use AI to design a website.