AI answers are a zero-sum surface: when a buyer asks "what's the best option for X", the engine names a handful of brands, and every name that isn't yours is a competitor's. Tracking your own mentions tells you half the story. Competitor tracking tells you the half that stings — who's being named instead of you, and whether the distance is closing.
Measured on the same answers — not parallel ones
The design principle: your brand and its competitors are measured on the identical prompts and the identical answers. When a sweep runs your panel against the four engines, the judge extracts every brand mention from each answer — yours and each tracked competitor's, with rank and framing. There is no separate competitor sweep and no separate competitor panel.
This matters because it makes comparisons airtight. Engines are non-deterministic; if your brand and a competitor were measured on different runs, any gap could be an artifact of sampling. Measured within the same answers, a gap is real: the engine composed one response, and it named them ahead of you — or didn't name you at all.
You choose which competitors to track when setting up your brand (Brandflare suggests them during onboarding), and your plan sets the ceiling: 2 tracked competitors on Starter, 5 on Trial, Pro, and Agency — exact limits in plans and limits.
The three competitive metrics
Share of voice. Of all brand mentions in your panel's answers — yours plus your tracked competitors' — the percentage that belong to you. Share of voice is the single best summary of who owns your category's AI answers: a brand can have a respectable mention rate and still hold a sliver of voice if competitors are named more often, in more answers.
Head-to-head mention rates. Your mention rate beside each competitor's, computed from the same runs. This is the direct comparison — on the questions that matter in your category, how often does each brand make it into the answer? Because the underlying answers are shared, you can always click through to the runs where a competitor appeared and you didn't, and read exactly how the engine framed the choice.
Gap trends. The competitor gap — your visibility minus theirs — trended over time. The absolute gap tells you where you stand; the trend tells you whether your work is closing it. A 20-point gap that's shrunk from 30 is a campaign that's working; a 5-point gap that's widening is an early warning worth acting on while it's still cheap to fix.
Reading competitive data well
- Rank is part of the picture. Two brands mentioned in the same answer aren't tied — the judge records each mention's position, and being consistently named first is a durably different outcome from being consistently named last. The prompt matrix shows where each of you lands, prompt by prompt.
- Engines disagree; use it. A competitor may dominate Perplexity's retrieval-heavy answers while you lead on Gemini. Per-engine breakdowns show where each rival's strength actually lives, which tells you what kind of footprint — coverage, citations, entity presence — is producing it.
- Watch for displacement, not just decline. A falling share of voice with a flat mention rate means a competitor is being added to answers, not that you're being dropped from them. The remedies differ, and only the head-to-head view distinguishes the two cases.
- Don't over-read one sweep. As everywhere in Brandflare, single-run differences are noise; sustained patterns across sweeps are signal.
Choosing whom to track
Track the brands buyers actually compare you against — which isn't always your strategic archrival. The free audit at /audit and the answers in your first sweeps will show you who engines actually name alongside you; if a brand you've never considered a competitor keeps appearing in your prompts' answers, that's the market telling you your tracked set is wrong. Competitor slots are editable, but remember that changing the set resets the meaning of your share-of-voice trend, just as changing prompts would reset your score's yardstick.
Where to go deeper
The method behind measuring rivals on one yardstick is covered in how do I track competitors in AI answers?, and the public catalog on Explore lets you see visibility scores for openly tracked brands in your category before you commit to tracking anyone.