What to do this week:

Run ten prompts yourself. Open ChatGPT, Gemini, and Perplexity. Ask the questions your customers would ask without using your brand name. Count how many times you appear out of 30 total responses. That rough number is your starting visibility score.

Write down who else appears. The competitors showing up in your answers are the ones winning the AI visibility race right now. Understanding who they are and why they're appearing is the first step toward closing the gap.

Do it again next month. One measurement is a snapshot. Two measurements, a month apart, is the beginning of a trend. That trend is what turns visibility tracking from a curiosity into a strategic advantage.

Boost Background:

Last quarter we onboarded a resort group that was convinced their AI presence was fine. Their marketing team had checked ChatGPT once, seen the property mentioned, and moved on. When we ran a proper audit, 50 prompts across five platforms, the brand appeared in just 7% of relevant answers. Their closest competitor? 34%.

That gap is what a visibility score measures. It's the percentage of relevant AI-generated answers that mention your brand. And it's quickly becoming the most important metric most marketing teams aren't tracking.

Why this metric exists

Traditional SEO gives you rankings, traffic, and click-through rates. None of those capture what's happening inside AI answers. When someone asks ChatGPT for the best boutique hotel in Napa or asks Perplexity which museums to visit in Chicago, your brand either appears in the response or it doesn't. There's no "page two." There's no ad position to buy (yet). You're in the answer or you're invisible.

A visibility score quantifies that. Run 50 prompts your ideal customer would ask. If your brand shows up in 15 of them, your visibility score is 30%. Your competitor might be at 55%. Now you have a number you can track, benchmark, and improve.

AirOps research found that only about 30% of brands maintain visibility in consecutive AI-generated answers, and just 1 in 5 sustain it across five consecutive runs. That instability is exactly why measuring at scale matters more than any single spot check.

What the research says about measuring it

This is where things get interesting. SparkToro and Gumshoe.ai published landmark research in January 2026 that tested whether AI recommendations are consistent enough to measure. They had 600 volunteers run 12 prompts through ChatGPT, Claude, and Google AI nearly 3,000 times. The finding: there's less than a 1-in-100 chance that any two runs of the same prompt will produce the same list of brands.

That sounds like it makes tracking pointless. It doesn't. It means you have to measure the right thing. While ranking position is essentially random, visibility percentage across many runs tells a real story. The same brands kept appearing across most responses even as their position shuffled. A follow-up analysis from Search Engine Land found that when ChatGPT was asked for B2B software recommendations 100 times, roughly five brands appeared in 80% or more of responses. Those brands had earned a durable association in the model's understanding.

This is the distinction we emphasize with every client: stop caring about whether you're listed first or third. Start caring about whether you appear at all, and how often.

How we calculate it

We build a prompt set of 25 to 50 questions a real customer would ask, without using the brand name. We run those across ChatGPT, Gemini, Perplexity, Claude, and Microsoft Copilot. For each response, we log whether the brand appears, how it's positioned, the sentiment, and which competitors show up.

The primary number is simple: brand appearances divided by total prompts, expressed as a percentage. But we also track supporting metrics. Sentiment tells you whether AI describes you favorably or with qualifiers like "budget option." Citation source analysis reveals where AI is getting its information about you. Share of answer shows how you compare against specific competitors across the same prompt set.

We run this monthly. A brand that moves from 12% to 28% over three months is building real momentum. A brand that drops from 25% to 15% has a problem worth investigating before it compounds.