AI share of voice vs. presence: what each metric counts
AI share of voice measures your share of the brand mentions in a set of captured answers. Presence measures how many of those answers mention you at least once. You need both to distinguish broad coverage from repeated mentions in a few answers.
The formulas use different denominators
In Braign's standard report, share of voice = your brand mention occurrences ÷ all included brand mention occurrences × 100.
Presence = included answers mentioning your brand at least once ÷ all included answers × 100.
For share of voice, “included” matters. The brand set can contain your brand, the competitors you entered, and discovered brands included in the view. Excluding a brand changes the denominator. Confirm the selection before comparing reports.
Presence counts a brand once within each answer. Ten repetitions of the same name in one response still contribute one answer toward that brand's presence. Answers with no recorded mention of your brand remain in the presence denominator when they are included in the report.
A brand can have 60% presence and 20% share of voice
Consider five fully analyzed answers and two included brands. This is an illustrative example.
| Answer | Your brand mentions | Competitor mentions |
|---|---|---|
| A | 1 | 3 |
| B | 2 | 4 |
| C | 1 | 3 |
| D | 0 | 4 |
| E | 0 | 2 |
| Total | 4 | 16 |
Your brand appears in A, B, and C: three of five answers, or 60% presence. It has four of the 20 brand mention occurrences, so its share of voice is 20%.
The competitor appears in every answer, giving it 100% presence and 80% share of voice. Presence percentages across brands can add to more than 100% because several brands can appear in one answer. Share-of-voice percentages divide a common pool of included mentions.
If a view instead counts each brand only once per answer, this example gives your brand three of eight brand-answer appearances, or 37.5% mention share. The counting rule belongs alongside the result. Braign's standard report displays “All occurrences” for its mention-share calculation.
What to do when the metrics disagree
High presence with low mention share gives you a specific review task: open answers where the brand appears briefly and competitors receive more discussion. Check whether those extra mentions contain useful recommendations or simply repeat a name. Mention volume alone does not establish preference.
Low presence with a larger mention share can come from a few answers that discuss the brand repeatedly. Find the questions where it never appears. Those omissions may point to a different category, use case, or buyer requirement.
Read “Positive when mentioned” and “Named first” beside these metrics, then inspect the original wording. An answer can name your brand early and warn against buying it. The report guide explains those additional denominators.
Keep the measurement conditions attached to the number
Record the question set, selected models, search settings, included brands, counting rule, run date, and completed-answer count. A 20% share in a price-focused question set cannot be read as the same measurement as 20% in a sustainability-focused set.
“AI visibility score” can mean different things across vendors. Ask for the formula before comparing a score with Braign's presence or mention share. For example, Promptwatch describes a score based on position, context, and competitors. That is a different measurement from either formula above.
Use the run-comparison guide when tracking the same report over time. If a model or question is missing from one run, establish which pairs are comparable before announcing a change.
See the evidence in a report
Explore the public sample, or choose your brand, buyer questions, and models to price your own report.