How to Measure AI Search Visibility: A Framework for Professional Services

Every post on this site tells you to optimise for AI search. None of them tell you how to know if it is working.

That is a problem. You cannot improve what you do not measure. And unlike traditional SEO, where Google Search Console gives you impressions, clicks, and positions for free, AI search has no standard measurement tool. There is no dashboard. There is no “AI Search Console.”

But that does not mean you cannot measure it. This is the framework we use in every AI Search Visibility Audit we deliver, adapted for firms that want to track it themselves.

Why measurement matters now

AI search traffic converts at 5x the rate of Google search. This is because AI queries tend to be high-intent: the person asking “which restructuring advisory firm should I hire?” is closer to a buying decision than someone Googling “restructuring advisory.”

But conversion only matters if you appear. Most professional services firms have a citation rate below 10%. They are invisible to nearly half their potential buyers.

The three-layer framework

We measure AI visibility across three layers. Each answers a different question.

Layer 1: Citation rate

This is the baseline metric. Of the queries a potential buyer might ask an AI model about your services, what percentage produce an answer that mentions your firm?

How to measure it:

1. Write 20 buyer-intent queries. These are the questions your prospective clients would ask an AI model. Examples for a management consultancy: “Best digital transformation consultancies in the UK”, “Which consulting firms specialise in AI strategy for mid-market companies?”, “Recommend a consultancy for operational restructuring.”

2. Run each query across at least two AI models. We recommend ChatGPT (GPT-4o) and Perplexity as the minimum. Add Claude and Gemini if your audit budget allows.

3. For each response, record: Was your firm mentioned? If yes, in what position (1st, 2nd, 3rd)? What was the exact quote?

4. Calculate your citation rate: (queries where you appeared) / (total queries x models tested).

Our free AI Citation Checker automates this process. Enter your domain, and it runs buyer-intent queries across ChatGPT and Perplexity, parses the responses for citations, and calculates your score.

Layer 2: Citation accuracy

Being cited is only valuable if the AI describes your firm correctly. We regularly see AI models attribute the wrong specialisms, locations, or capabilities to firms. A law firm specialising in employment law being recommended for corporate M&A is worse than not being recommended at all.

How to measure it:

For every citation found in Layer 1, score it:

• Correct: The AI accurately describes your specialism, location, and capability.
• Partially correct: The firm is mentioned but the description is inaccurate or outdated.
• Incorrect: The AI attributes services you don’t offer or confuses you with another firm.

Track the accuracy rate: (correct citations) / (total citations).

What to do about inaccuracies:

Inaccurate citations usually stem from conflicting information across the web. If your website says one thing, your LinkedIn says another, and an old press article says a third, the AI model will average across them, often incorrectly.

The fix is consistency: ensure your service descriptions, specialisms, and positioning are identical across your website, LinkedIn company page, directory listings, and llms.txt file.

Layer 3: Citation conversion

This is the hardest layer to measure and the most commercially important. When someone sees your firm recommended by an AI model, do they visit your site? Do they enquire?

How to measure it:

1. Direct referral tracking. Some AI platforms pass referral data. In GA4, check Acquisition > Traffic Acquisition for referrers containing “perplexity”, “chatgpt”, “bing” (Copilot), or “gemini”. ChatGPT and Claude traffic is harder to attribute because they do not consistently pass referral headers.

2. Dedicated landing pages. Create a landing page specifically for AI-referred visitors (e.g., “/ai-recommended”). Mention this URL in your llms.txt and structured data. Any traffic to this page is likely AI-referred.

3. Intake form tracking. Add “How did you hear about us?” to your enquiry form with “AI assistant (ChatGPT, Perplexity, etc.)” as an option. This is low-tech but surprisingly effective. Several of our audit clients have added this and found 10-15% of new enquiries now select it.

4. Brand search uplift. Monitor branded search volume in Google Search Console. An increase in people Googling your firm name often correlates with increased AI visibility, because people see you recommended by an AI and then search for you directly.

The 20-query monthly audit

Here is the exact process we recommend for a monthly self-audit. Total time: 45-60 minutes.

Step	Time	What You Do
1. Update query list	10 min	Review and refresh your 20 buyer-intent queries. Add any new service lines or market changes.
2. Run queries	20 min	Test each query in ChatGPT and Perplexity. Record citations, positions, and quotes.
3. Score accuracy	10 min	For each citation, mark correct / partially correct / incorrect.
4. Check analytics	5 min	Review GA4 for AI referral traffic. Check intake form responses.
5. Update tracker	5 min	Log results in your monthly tracking spreadsheet. Note changes from last month.

Over time, this monthly data gives you a trend line. Are you becoming more visible or less? Which queries are improving? Which models cite you more? This trend data is what informs your earned media strategy and on-site optimisation priorities.

Benchmarks for professional services firms

Based on audits we have delivered across management consultancies, law firms, accounting practices, and executive search firms:

Top quartile

35%

Average

12%

Bottom quartile

3%

• Top quartile (25-40% citation rate): Firms with active content programmes, earned media, structured websites, and schema markup. These firms appear in 1 in 3 relevant AI queries.
• Average (8-15%): Firms with decent websites but no deliberate AI optimisation. Occasional citations, usually for their most well-known service line.
• Bottom quartile (0-5%): Firms with gated content, vague service descriptions, no schema markup, and minimal third-party mentions. Effectively invisible.

The gap between top and bottom quartile is widening as early movers build compounding advantages.

Common measurement mistakes

Testing with the wrong queries. If your queries are too generic (“best consulting firm”), you are measuring brand awareness, not service visibility. Use specific buyer queries that match your actual service lines.

Testing once and stopping. A single measurement is a snapshot. AI models update continuously. A monthly cadence is the minimum for tracking trends.

Ignoring accuracy. A 30% citation rate means nothing if half the citations describe your firm incorrectly. Accuracy measurement catches positioning problems before they cost you business.

Confusing AI traffic with AI visibility. Traffic is a lagging indicator. Many people who see your firm recommended by an AI will Google you directly rather than clicking a link. Measure citation rate as the leading indicator and traffic as the trailing one.

Tools that help

The measurement space is evolving rapidly. A few tools worth evaluating:

• BriefingHQ Citation Checker (free): Automated citation rate scanning across ChatGPT and Perplexity. Good for a quick baseline.
• Otterly.ai: Tracks AI mentions across multiple models with historical data.
• SE Ranking AEO module: Integrates AI visibility tracking with traditional SEO monitoring.
• Manual audit: Still the most flexible approach for professional services firms with specific, niche queries.

Start measuring

The first step is always a baseline. Run a free citation scan to see where you stand today. From there, the monthly 20-query audit gives you the trend data you need to make informed decisions about where to invest your earned media and on-site optimisation effort.

Want the full picture? Our AI Search Visibility Audit runs this across all four major AI models with competitor benchmarking and accuracy scoring built in.