Stage 1 · the AEO audit

AI Visibility Audit

The AI Visibility Audit shows whether AI assistants put your company on the shortlist when your buyers ask, and what to do about it. We score your buyers' prompts across five assistants, check the technical, on-site and off-site signals they rely on, and hand you a phased 90-day plan.

A fixed fee, confirmed on the call · ten working days · credited in full against the Implementation Programme. Not ready for an audit? The free AI access check tells you in about a minute whether AI assistants can read your site at all.

Why an audit first

Why start with an audit?

Because AI decides the shortlist before anyone visits your site, and you currently can't see it. 82% of B2B software buyers acted on a recommendation from an AI assistant in the last two years, and review sites and AI chatbots now shape the shortlist in equal measure (G2, July 2026).

Each assistant builds its answer differently. 91% of citations appear in only one of ChatGPT, Perplexity or Google's AI Overviews (Growth Memo, H1 2026). Checking one assistant tells you little about the rest.

The usual tools don't show it either. Search Console's Generative AI report (since 31 August 2026) shows impressions in AI Overviews and AI Mode combined, with no clicks and no queries (Google). 45% of marketing leaders say they cannot measure AI visibility (Semrush, June 2026).

How the audit earns your trust

  • Every finding about your company is checked live by a person and logged in a verification log.
  • Every external claim carries its source.
  • If a finding changes during the work, the report says so.
  • Scores come from a fixed panel, not a single screenshot.
  • A “what not to spend money on” section, so you don't pay for tactics without evidence.
  • A plan your own team or SEO agency can run.

What you get

What's in the audit?

In one sentence

The AI Visibility Audit is a scored, sourced baseline of how AI assistants describe and recommend your company today, with a phased 90-day plan to improve it.

Eleven parts, in the same order every time, so you can compare the before and after directly.

30 / 50buyer prompts, standard or extended, agreed with you before we run them.
5AI assistants: ChatGPT, Google AI Overviews and AI Mode, Microsoft Copilot, Perplexity and Claude.
10working days to the finished report, with a readout call included.
01

Executive summary

Where you stand across your territories (head terms, niche terms, new categories), the thesis in one line, and the workstreams that follow.

02

Why it matters in your sector

How buyers in your sector use AI, and which assistants they actually open.

03

How assistants choose sources

How assistants search, what they read and what they cite, in plain terms. Updated quarterly.

04

Technical baseline

A live crawler access test across about ten answer and training bots, scored separately from the training bots, plus robots rules, rendering, sitemap completeness, duplicates and CDN rules.

05

On-site baseline

Machine identity (what your first heading and first 100 words say you are, with the title and description checked as hygiene), answer-content gaps, and proof you already have but don't use.

06

Off-site baseline

The roundups and listicles AI cites, review platforms, YouTube, LinkedIn, trade press and community threads such as Reddit, with Wikipedia and Wikidata where they apply. Product feeds for retailers, and branch listings for multi-location businesses.

07

Baseline prompt panel

Your buyers' prompts in five groups, scored 0 to 3 for each of five assistants: presence rate, mean score, zero-score prompts and competitor share of voice, with mentions and citations kept apart.

08

What not to spend money on

The tactics we'd skip in your case, and why, with sources.

09

The 90-day plan

Phased actions, each with an owner, an effort estimate and the metric it feeds.

10

How we'll measure it

The metrics, sources and cadence that will show whether the plan worked.

11

Appendices

The full prompt panel, the verification log and every source.

Plus

The readout call

We walk your team through the findings and the plan, and answer questions before you decide anything.

What a finished audit looks like

Findings are presented as tables and short verdicts, each tied to an action in the plan.

The format comes from live work. We're working with a Waterford-based B2B SaaS company on a 90-day AEO programme, with a public before-and-after case study planned.

Research is AI-assisted and human-verified. The verification log shows what we checked, how and when, so every claim about your company can be traced.

The prompt panel

How is the prompt panel scored?

We agree 30 buyer prompts with you (50 on the extended audit), run each one across five assistants, and score every answer from 0 to 3.

ScoreMeaningIn the answer
0AbsentYou don't appear.
1Mentioned, not citedYou're named in the text, with no link to your site.
2Cited as an optionYou're named and linked, listed among several vendors.
3Cited first or recommendedYou're the recommended option, or first in an ordered list.

Five prompt groups

  • Head terms: the category questions, such as “best [category] software for [use]”.
  • Niche terms: the specialist questions you should already own.
  • New category: the emerging category you want to enter.
  • Competitor: “alternatives to” and “X vs Y” questions.
  • Brand: what AI says when asked about you directly, and whether it is accurate: your category, your prices and your competitors.

Presence rate

The share of answers that name you at all, scoring 1 or more.

Mean score

Your average score out of 3, overall and for each prompt group.

Zero-score prompts

The prompts where you're absent, with who wins them instead and why.

Share of voice

Your mentions as a share of all vendor mentions in the same answers.

Every run uses a fresh logged-out session, the same wording and a location set to your market. We record the date and model version with every answer and archive the raw text, so every score can be traced back to what the assistant actually said.

The assistants

Which AI assistants do you test?

The major assistants your buyers use to research and compare. We test all of them because each one builds its answer from different sources. Google's AI features draw on Google's index, and Copilot draws on Bing's. ChatGPT relied on Bing's index until September 2026, and now searches OpenAI's own. Perplexity and Claude run their own indexes too.

Whatever the underlying index can't see, the assistant can't cite. That is why the technical baseline comes before any content advice.

ChatGPTGoogle AI OverviewsGoogle AI ModeMicrosoft CopilotPerplexityClaude

Google AI Overviews and AI Mode are scored together as one assistant, so the panel covers five.

Who it's for

Who is the audit for?

Heads of marketing, heads of digital and ecommerce, and scale-up founders at established companies whose buyers ask AI before they choose.

B2B

Software and services

Firms whose buyers ask AI for a vendor shortlist, where review platforms, roundups and comparison pages decide who gets named.

Construction

Building products

Manufacturers and independent merchants whose technical data, rebrands and crawler rules decide what AI can find and repeat.

Retail

Online pharmacy, health and beauty

Retailers where product feeds shape AI shopping answers, with a compliance step built into the plan.

The audit is built for established companies. For a quick look at the basics first, the free AI access check is the lighter first step.

How it works

What happens after you book?

01 · Agree the panel

Kick-off

We agree your competitors, your territories and the prompt list. We'll ask for your products or services, the review platforms you use, your feed status if you sell online, and read access to Search Console and GA4 if you can give it.

02 · Run and verify

Research

We run the panel, test crawler access live, and research your on-site and off-site footprint. Every finding is checked by a person and logged.

03 · Readout

Report and call

You get the report and a readout call that walks through the findings and the 90-day plan. Ten working days in total.

Scope

Standard or extended?

Both follow the same eleven-part format and take ten working days, and both are a fixed one-off fee. The extended audit covers more ground. We agree which one fits, and tell you the fee, on a 20-minute scoping call.

Standard

One brand, one market

30 buyer prompts across five assistants. The right fit for most single-brand B2B companies.

30 buyer prompts
Extended

Multi-brand, Ireland and UK, or ecommerce

50 buyer prompts across five assistants. For several brands, for Ireland and the UK together, or for ecommerce, where it includes a review of your product feeds.

50 buyer prompts
Fully credited

If you start the AEO Implementation Programme within 60 days, the full audit fee comes off the programme fee. So the audit costs you nothing if you go ahead. See what each stage includes.

FAQ

Questions about the audit

What is an AI Visibility Audit?

It's a scored, sourced baseline of how AI assistants describe and recommend your company today, with a phased 90-day plan to improve it. We score your buyers' prompts across ChatGPT, Google's AI Overviews and AI Mode, Copilot, Perplexity and Claude, and check the technical, on-site and off-site signals those assistants rely on.

How long does the audit take?

Ten working days. It ends with a readout call where we walk you through the findings and the plan.

How much does it cost, and how does the credit work?

The audit is a fixed one-off fee, excluding VAT, in two sizes: standard (one brand, one market, 30 prompts) and extended (multi-brand, Ireland and UK, or ecommerce with feed review, 50 prompts). We confirm which one fits and what it costs on a 20-minute scoping call, and in writing before you commit. If you sign up for the AEO Implementation Programme within 60 days, the full audit fee is credited against it.

Who chooses the prompts?

We draft the panel from your categories, your competitors and the questions your sales team hears. You agree it before we run anything, and the final list is included in the report.

What do you need from us?

Your main competitors, your products or services, the review platforms you use and, if you sell online, the state of your product feeds. Read access to Search Console and GA4 lets us include your first-party AI data.

How is this different from the free AI access check?

The free check is an automated test, taking about a minute, of whether AI assistants can read your site at all. The audit is ten working days of research, checked by a person. It shows what the assistants actually say about you and your competitors, why, and what to do about it.

Can we act on the audit without you?

Yes. Every action in the 90-day plan has an owner, an effort estimate and the metric it feeds, so your own team or your SEO agency can run it. If you'd rather we did the work, the audit fee is credited in full against the Implementation Programme.

A fixed fee · credited in full against the programme

Find out if AI puts you on the shortlist.

Your buyers' prompts scored across five AI assistants, and a phased 90-day plan, in ten working days.