Marketing has always had a metric called share of voice: your slice of the total advertising noise in your category. It was measured in ad spend and media mentions, and it predicted market share well enough that brands managed it seriously. AI assistants have quietly created a new version of the same metric, and most padel brands have never seen theirs.

AI share of voice is the percentage of recommendation mentions your brand captures when people ask AI assistants the buying questions in your category. Ask the assistants “which padel racket should an intermediate player buy” a hundred times and count every brand named: if your brand appears in fifteen of the answer slots, your share of voice for that question is fifteen percent. Aggregate across the questions that matter in a market and you have a single number describing your position in the channel where buying decisions increasingly start.

How the measurement actually works

Rigour matters, because assistants are probabilistic: the same question asked twice can produce different answers. A credible measurement handles that by repetition and breadth.

The method we use in our published research takes a fixed panel of realistic buyer questions for one market, puts each question to each major AI platform many times, and records every brand mention with its position in the answer. Repetition smooths the randomness; breadth across questions prevents one flattering prompt from distorting the picture; splitting results per platform preserves differences the average hides. Our first public edition, the padel racket report for Spain, shows the output: a leaderboard of brands by share of AI mentions, measured across OpenAI, Perplexity, and Gemini.

The per-platform split deserves emphasis. Assistants retrieve from different sources, so a brand can hold a strong position on one platform and be absent from another. Those are different problems with different fixes, and one blended number would hide both.

What the number is good for

Share of voice turns an anxiety into a management object. “Are we visible in AI” is not actionable; “we hold nine percent against the leader’s thirty” is. It gives brand teams a baseline, a competitive frame, and a way to see movement.

It also localises the problem. Share of voice is measured per market, and the same brand can dominate in one country and barely register in another, because reviews, retailers, and press differ by market. A brand strong in Spain may be invisible in the answers shaping the fast-growing markets we track, from Milan to London. Expansion plans should read those numbers first.

And it moves. When a brand lands in the right comparison articles, fixes its product data, and builds review coverage, its mention count climbs in the following measurement rounds. Watching the number respond to work is what turns AI visibility from a mystery into a channel.

What moves it

The inputs are unglamorous. Coverage in the review and comparison content assistants read, because that is where product recommendations are grounded. Clean, crawlable product and range pages that let machines quote specs. Consistent brand naming across your site, retailers, and press. Content that answers need-specific questions (control versus power, beginner versus advanced), because those specific questions are where the podium reshuffles. Sponsorships and pro contracts help indirectly, by generating written coverage; the coverage is the input, machines cannot read a shirt.

Where to start

Get your number. The free Brand AI Visibility Checker reads our published report data and shows a brand’s share of voice, rank, and gap to the category leader in seconds. If your brand is in the data, you will know your position; if it is absent, that is the finding.

Then treat it like any other KPI: baseline, fixes, re-measure. Our equipment brands page lays out the full loop.

The uncomfortable version of this article is one sentence long: the assistants are already publishing a ranking of your category, your buyers are already reading it, and your only choice is whether you measure your position or ignore it.

AI visibility