AfieldLeave the audit with an edit you can make
Afield reads your public site and checks titles, descriptions, headings, images, links, sitemap, robots rules, structured data, social previews and available speed measurements. The dedicated audit scans up to eight pages per run. Findings name the affected URLs and show what needs attention.
Choose a finding to generate a proposed fix, or prepare a task for your connected GitHub repository. You can start the public audit before connecting Search Console or a CMS.
What a useful finding contains
The example below demonstrates an audit handoff; it is not a scan result or a customer case study.
| Finding | Affected page | Proposed change |
|---|---|---|
| Generic page title | /features/field-reports | Replace “Platform” with “Field reports for restoration jobs” if the page documents that workflow. |
| A guide link returns 404 | /docs/setup → /docs/old-import | Link directly to the current import instructions; preserve a redirect if the old guide moved. |
| No product example in the opening | /features/field-reports | Show a report with room photos, readings and reviewer notes before asking for a demo. |
Each row is a different assignment. A broken link needs a destination fix; an unclear product page needs a product example. A score alone does not tell an editor which work to do.
Four different questions, four evidence sources
| Your question | Afield can use | Result scope |
|---|---|---|
| Can search engines read this page? | Public HTML, rendered content and technical checks | The pages actually read in this crawl. |
| Which searches bring impressions and clicks? | Your connected Google Search Console property | The selected dates, pages and queries. |
| Where does this query appear now? | A public Google organic-results snapshot when available | The returned top-ten results for the specified market, not an all-keyword rank tracker. |
| Does an AI answer mention my brand? | An answer check using configured models or answer APIs | The tested questions, actual models or API methods, and available answers; this is not a consumer-app ranking. |
Inspect the actual model answer
Afield has two scoped answer checks. A fixed-question baseline submits one exact buyer question to the answer APIs configured for that installation. A deep GEO simulation uses a generated buyer-question matrix across configured models; it needs successful evidence from at least six buyer angles and two models before reporting a deep GEO conclusion.
Availability depends on the configured providers and successful responses. Keep the question, actual model or API method, date and original answer together. Inspect brand mention, description accuracy, recommendation evidence and returned source references separately. A failed answer is missing evidence.
These samples do not establish a ranking across consumer AI applications. For a comparison, use the same question wording and provider/model scope; record any change. Choose the answer-check scope and use the review worksheet. Use Search Console separately for observed Google queries and clicks.
Review the replacement before it reaches the site
For a copy issue, request the exact title, opening or replacement section. For a technical issue, carry the affected URL and evidence into the GitHub fix request. For an article, use the connected publishing destination and choose draft or publish.
After applying the change, open the public page and repeat the relevant check. Publishing a corrected page and Google indexing it are separate events. Search Console is the place to inspect indexing and later query performance.
Run a website audit · Connect Search Console · Prepare a repository fix