Fact consistency prompt

A fact consistency prompt reads every page and lists each fact that appears in more than one place, such as name, address, numbers, services and dates, and flags any that disagree. Conflicting facts make AI tools hedge, or describe you wrongly.

Stage
Check, step 33 of 40
Use it
Before launch, after every big edit, and once a quarter.
You need
  • The text of your pages
  • Your Google Business Profile, LinkedIn and directory listings
  • Your fact list
You get
A list of repeated facts, every version of each, and which one is correct. Output schema included.

The prompt

Read all the pages below. List every fact about the business
that appears more than once: name, address, phone, years in business,
client counts, services, prices, team members and dates.
For each fact, show every version and the page it's on.
Flag any fact with more than one version.
Then compare with my fact list and mark which version is correct.
[paste page text, profile text or sitemap URLs]
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": "consistency-check",
  "name": "Fact consistency prompt",
  "stage": "check",
  "step": 33,
  "of": 40,
  "version": "2026-11-19",
  "url": "https://websitesdesigner.com/insights/prompts/consistency-check/",
  "inputs": {
    "paste_page_text": "[paste page text, profile text or sitemap URLs]"
  },
  "instructions": [
    "Read all the pages below.",
    "List every fact about the business that appears more than once: name, address, phone, years in business, client counts, services, prices, team members and dates.",
    "For each fact, show every version and the page it's on.",
    "Flag any fact with more than one version.",
    "Then compare with my fact list and mark which version is correct.",
    "{{paste_page_text}}"
  ],
  "output": "A list of repeated facts, every version of each, and which one is correct.",
  "checks": [
    "Every conflict names the pages involved.",
    "Off-site profiles are included.",
    "Each conflict has a correct version marked."
  ],
  "output_schema": {
    "type": "object",
    "properties": {
      "facts": {
        "type": "array",
        "items": {
          "type": "object",
          "properties": {
            "fact": {
              "type": "string"
            },
            "versions": {
              "type": "array",
              "items": {
                "type": "object",
                "properties": {
                  "value": {
                    "type": "string"
                  },
                  "where": {
                    "type": "string"
                  }
                },
                "required": [
                  "value",
                  "where"
                ],
                "additionalProperties": false
              }
            },
            "conflict": {
              "type": "boolean"
            },
            "correct_value": {
              "type": [
                "string",
                "null"
              ]
            }
          },
          "required": [
            "fact",
            "versions",
            "conflict",
            "correct_value"
          ],
          "additionalProperties": false
        }
      }
    },
    "required": [
      "facts"
    ],
    "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": {
    "facts": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "fact": {
            "type": "string"
          },
          "versions": {
            "type": "array",
            "items": {
              "type": "object",
              "properties": {
                "value": {
                  "type": "string"
                },
                "where": {
                  "type": "string"
                }
              },
              "required": [
                "value",
                "where"
              ],
              "additionalProperties": false
            }
          },
          "conflict": {
            "type": "boolean"
          },
          "correct_value": {
            "type": [
              "string",
              "null"
            ]
          }
        },
        "required": [
          "fact",
          "versions",
          "conflict",
          "correct_value"
        ],
        "additionalProperties": false
      }
    }
  },
  "required": [
    "facts"
  ],
  "additionalProperties": false
}

What good output looks like

  • Every conflict names the pages involved.
  • Off-site profiles are included.
  • Each conflict has a correct version marked.

What experts miss

  • Check off-site profiles too. Your Google profile, LinkedIn and directory listings feed AI answers as well as your site.
  • Write the business name one way. "Website's Designer", "Websites Designer" and "WD" can read as three different businesses.
  • Schema counts as a page. Paste your JSON-LD too, because it often holds an old address or phone number.

Why it works

AI tools build their picture of a business from many pages and profiles. When the facts disagree, the safest answer for the tool is a vague one, or none. One consistent set of facts is the base of your entity profile.

How to adapt it

  • Add your invoices and email signature, which often carry old details.
  • Run it per language version.

Mistakes to avoid

  • Fixing conflicts on the site and forgetting the profiles.
  • Letting a rounded number creep in, such as 10+ years on one page and 12 on another.

Go deeper

See entity profile and content QA. The fix starts with the entity profile prompt. This prompt is part of the check stage. It follows the AI website design prompts guide and the six steps in how to use AI to design a website.

Read the prompts guideAll 40 prompts

Related prompts

Entity profile prompt

An entity profile prompt writes one canonical description of your business: name, category, what you do, for whom, where and since when, plus the profiles that are really yours. You copy it word for word to your site, your Google profile and directories.

Get the prompt

Fact list prompt

A fact list prompt gives the AI a numbered list of true facts and forbids anything else. Each claim in the copy carries its fact number, and gaps get marked instead of filled with invented numbers or testimonials.

Get the prompt

Schema markup prompt

A schema markup prompt asks the AI to write JSON-LD structured data for a page, using only facts visible on that page. It returns one linked graph and a list showing where each property comes from, so you can check it.

Get the prompt

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.

Get the prompt

Want this done for you?

Want to know what the web says about your business? Book a call and I'll run the check with you.

Book a 20-minute call