Our GEO Audit Template for 2026

A GEO audit is the first thing we run on a new account. It is not a slide deck. It is a checklist an auditor can follow, with a deliverable at the end that says what is broken, what is missing, and what to do first. This is the template we use in 2026, written for the operations lead or marketing director who has to commission one and read the result.

We run the audit on Promptwatch, the platform we run client programs on. Each step below names the Promptwatch feature that produces the evidence, so the audit is reproducible rather than a set of opinions. The template has six sections. An auditor who skips a section produces a partial audit, and partial audits are how clients end up fixing content when the actual problem is the crawl.

1. Prompt baseline

The audit starts with the prompt list. We do not accept a prompt list the client hands us without checking it, and we do not invent one. We build it from three sources: the client's own search console data, sales call notes from the team that talks to buyers, and the query fanouts report, which accounts for the long, multi-intent prompts real people now type.

We use prompt tracking in Promptwatch, which gives us search volumes, difficulty scores, and fan-outs per prompt, plus country, state, and city tracking where the client has local intent. The deliverable is a frozen prompt list, tagged by type: informational, navigational, commercial, transactional. The list is the baseline. Every later section of the audit is measured against it. If the prompt list drifts, the audit is unreadable, so we lock it in week one and add new prompts as a separate tracked cohort.

The audit answers two questions here. Which prompts does the client care about. And on which of those prompts are they currently visible, absent, or losing ground. That is the scope. Everything else in the audit is diagnosis.

2. Citation position

With the prompt list frozen, we measure where the client is cited and where they are not. Citation analytics in Promptwatch gives us page-level and domain-level citations, plus Reddit citations, YouTube citations, and offsite mentions, with citation type breakdowns. We use citation trends to see how citation sources and volume move over time, not just a point-in-time count.

The deliverable is a prompt-by-prompt map. For each prompt, we record whether the client is cited, which page is cited, which competitor is cited instead, and what kind of citation it is: a link, a passing mention, a quote, a list inclusion. The kind matters. A link citation sends traffic. A passing mention often does not. The audit distinguishes them because the fix differs.

We also pull the citation top pages report, which shows which of the client's pages actually earn citations. The pattern is usually uneven. A small set of pages carries most of the citations, and a large set of pages earns none. That tells us where the citeable content already lives and where it does not.

3. Crawler logs

This is the section most audits skip, and it is the section that finds the real problems. Prompt tracking tells you whether you got cited. It does not tell you why. The why lives in the crawler logs.

Promptwatch Agent Analytics records the AI crawlers that hit the site, ChatGPTBot, ClaudeBot, PerplexityBot, GoogleOther, and the Meta AI crawler, and traces the crawl-to-citation path. We pull the crawler trend to see crawl frequency over time, and the top crawler pages to see which pages the engines actually read.

The deliverable is a list of crawl problems. A page that is cited but hit once a week and returning 5xx on half the hits. A page that should be cited but is blocked in robots.txt. A crawler that stopped hitting the site the week a CDN rule changed. A page that is crawled but never cited, which is a content problem, not a crawl problem. Each finding gets a specific fix, not a recommendation to "create more content."

We also use the logs to push back on clients who block AI crawlers wholesale. Blocking a crawler to save bandwidth is a reasonable choice. Blocking it and then expecting to be cited by that engine is not. The logs let us show the client exactly which crawlers they are blocking and what the citation trend did after the block.

4. Visitor analytics and conversions

Visibility without pipeline is a billboard. The audit has to close the loop. Promptwatch visitor analytics uses a lightweight script or a GTM template to track AI-referred traffic and conversions. We pull the visitor total, the visitor trend, the top pages, the top sources, and the top locations.

The deliverable is an attribution read. We compare the citation map from section 2 to the visitor trend from section 4. A prompt where the client is cited with a link and sees traffic is a working citation. A prompt where the client is cited with a passing mention and sees no traffic is a visibility impression without pipeline. A prompt where the client sees traffic without a citation is a referral the engine sent without naming the brand. All three happen, and the audit names which is which.

We do not promise a conversion number in the audit. We promise a measurement. The first quarter is usually baseline collection. The audit sets the baseline. The program reports against it.

5. Content gap analysis

With the prompt map and the crawl read in hand, we know which prompts the client is absent on and which competitors are cited instead. Content gap analysis in Promptwatch surfaces the prompts where the client is absent and a competitor is cited, and proposes a draft to fill the gap. We use the content gap stats for volume and the content gap recommendations for the prioritized list.

The deliverable is a prioritized content list, not a content calendar. For each gap, we record the prompt, the competitor currently cited, the page that would need to exist or be fixed, and whether the gap is a content gap or a crawl gap. A prompt where the client has a relevant page that is not being crawled is a crawl gap, and the fix is in section 3. A prompt where the client has no relevant page is a content gap, and the fix is a draft.

We are strict about the distinction. Writing content to fix a crawl problem is the most common waste in a GEO program. The audit separates the two so the program does not spend on the wrong one.

6. Action board and reporting

The audit ends with a prioritized action list, not a strategy document. Promptwatch Unified Actions produces an AI-generated to-do list from the visibility data, and we use the action board to rank the findings by impact and effort. Each action names the section it came from, the fix, and the metric we will watch to confirm the fix worked.

The deliverable is a one-page action board the client can sign off on, with the underlying rows attached. Crawl fixes first, because they are cheap and usually move the trend. Content gaps second, prioritized by prompt value and competitor weakness. Measurement and reporting third, because the program has to report against the baseline the audit just set.

What the audit is not

The audit is not a content strategy. It is not a brand strategy. It is not a guarantee of outcomes. It is a diagnosis of the current state of the site in AI search, with a prioritized fix list and a measurement baseline. A client who wants a strategy document should commission one separately. A client who wants a guarantee should buy an ad.

What we will not do

We will not run an audit without crawler logs. We will not write a content list before we have the crawl read. We will not promise a citation lift in the audit. We will not invent a case study result to close the engagement. The commitments we keep are the boring ones: no long-term contracts, senior specialists only, and an audit with the rows attached.

If you want the template above run on your site, with the prompt baseline, the crawler logs, the citation map, and a prioritized action board at the end, write to us at hello@1001seomedia.com. We will scope the prompt list, pull the logs, and show you what a 2026 GEO audit looks like when it is a checklist and not a slide deck.