See llms.txt for all machine-readable content.

Back to Templates

Audit page schema and AI citation signals with GPT-4o-mini and Google Sheets

Created by

Created by: Sara Soleymani || sarasoleymani
Sara Soleymani

Last update

Last update 8 days ago

Categories

Share

Quick overview

This workflow reads page URLs from Google Sheets, fetches each page to extract schema and SEO freshness signals, uses OpenAI GPT-4o-mini to assess entity consistency and generate recommended JSON-LD, then writes an audit row with scores and prioritized gaps back to Google Sheets.

How it works

  1. Runs manually to start an audit of all rows in the Google Sheets pages tab.
  2. Fetches each page_url over HTTP and parses the HTML to extract JSON-LD schema types, dates, author signals, canonical/robots tags, title/H1 query match, and a short content excerpt.
  3. Sends the extracted signals and excerpt to OpenAI (gpt-4o-mini) to score entity consistency, rank the biggest citation-impact gaps, and (when needed) generate a paste-ready JSON-LD block using placeholders for unknown values.
  4. Combines a deterministic technical checklist score with the OpenAI entity score into a weighted overall audit score and formats the top gaps and issues.
  5. Appends the full audit result (including scores, gaps, and recommended JSON-LD) to the Google Sheets audit tab with the run date.

Setup

  1. Create an OpenAI HTTP Header Auth credential with Authorization: Bearer <YOUR_API_KEY> for calls to the Chat Completions API.
  2. Add a Google Sheets OAuth2 credential and set the Google Sheets document ID in both the read (pages) and append (audit) steps.
  3. Create a spreadsheet with a pages tab containing at least page_id, brand, target_query, and page_url columns, and an audit tab with columns to capture the appended audit fields (for example: run_date, scores, top_gaps, recommended_jsonld, and issues).

Requirements

  • OpenAI account for the entity assessment and JSON-LD generation (gpt-4o-mini)
  • Google Sheets for page input and audit logging

Customization

  • Add a second fetch of the domain root to check homepage Organization schema, sameAs profile links, and llms.txt.
  • Adjust the technical checklist weights in the Parse & Score Audit node to match your own prioritization.
  • Pair it with a content citability scorer: when two pages score equally on prose quality but cite unequally, this audit usually explains the gap.