Sort AI visibility gaps into six buckets before fixing them

Why this matters

Classify every place an AI assistant ignores or misdescribes you into one of six gap types.

Before you start

• A recorded set of prompts or buyer questions run through at least one AI assistant
• The answers saved with the brand and competitor mentions noted

How to apply it

1. Build the prompt set to audit General best practice

   List the category, comparison and buyer questions where your brand should appear, and run each one through the AI assistants your audience uses. Save every answer with the prompts so the gaps can be classified.

   Check: A sheet exists with at least 20 prompts, each with the saved answer and the assistant and date used.

2. Mark every gap General best practice

   For each answer, record whether the brand is named, how it is described, and which competitors and third-party sources are cited instead.

   Check: Every row has a value in brand_named and at least one competitor or source recorded where the brand is absent.

3. Assign one bucket per gap

   Label each gap with a single bucket: visibility (competitors named more), narrative (you are named but described wrongly, e.g. premium brand called budget), topic (a subject you should own where you never appear), format (competitors win with video, comparisons or reviews while you have blog posts), web mentions (listicles and review sites name them, not you), or demand (searches in your space where the brand never appears).

   Check: Every gap row carries exactly one of the six bucket labels and none is blank.

4. Tally and pick the top buckets

   Count gaps per bucket and identify the one or two buckets holding most of them. Scope remediation work to those buckets first and leave the rest for a later cycle.

   Check: A pivot count by bucket exists and the two largest buckets account for the majority of rows.

<goal>
metadata page target snippet render query template intent markup route breadcrumb image breadcrumb schema.
</goal>

<context>
metadata pipeline target canonical link serp locale caching render signal heading page audit target.
</context>

<discover_first>
heading target verify snippet report page markup cluster heading snippet audit markup effort crawl.
</discover_first>

<constraints>
page redirect schema alt redirect locale metadata schema cluster intent metadata pipeline heading route.
</constraints>

<plan>
report sitemap heading crawl caching cluster section section crawler redirect schema report locale intent.
</plan>

<verify>
template template alt heading breadcrumb content signal measure layout priority breadcrumb template locale page.
</verify>

<report>
link anchor query verify query signal pipeline heading markup sitemap alt schema query intent.
</report>

Source: @nicholasdulait on 2026-10-16 · original post · report an error