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. Competitors 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 same questions and the same answers. When a check puts your questions to the five engines, the judge pulls every brand mention out of each answer, yours and each competitor's, with its position and how it was described. There is no separate competitor check and no separate set of competitor questions.
This makes comparisons airtight. Engines give different answers from one run to the next; if you and a competitor were measured on different runs, any gap could be an accident of sampling. Measured within the same answers, a gap is real: the engine wrote one response, and it named them ahead of you, or didn't name you at all.
Your competitors are found when the brand is set up, as many as your plan compares: two on Starter, five on Plus, Pro and Agency (plans and limits). You can change them at any time in Brand Settings, on the Tracking tab. Only questions that don't name your brand count, and a rival written with a note in brackets, such as FAB (First Abu Dhabi Bank), is matched under the plain name an answer uses.
What the Competitors section shows
Who AI names. A ranked list of you and your competitors by share of mentions, with logos. It matches Share of voice on your overview, and Who else gets named there.
Share over time. A chart where every brand has its own colour, with a sentence on the gap to your nearest rival. The size of the 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 cheap to fix.
Head-to-head. A table in plain words, one row per competitor: They usually lead, Neck and neck, You usually lead. It shows who's named first when you both appear, each competitor's usual place in the answer, and how AI describes them.
Share of voice, precisely
Share of voice is your share of all mentions among you and the competitors listed on your brand. Other companies an answer happens to name don't dilute it, so the figure means the same thing on your workspace and on a public profile. It's the best single summary of who owns your category's AI answers: a brand can be named respectably often and still hold a sliver of voice if rivals are named more, in more answers.
Your overview shows it with your place among the brands named, and it feeds Prominence in the Flare score.
Reading competitive data well
- Position is part of the picture. Two brands named in the same answer aren't tied. Being consistently named first is a different outcome from being consistently named last, and Questions & Prompts shows where you land, question by question.
- Engines disagree; use it. A competitor may dominate Perplexity and Google Search, which lean on live web results, while you lead on Gemini. Where a rival's strength lives tells you what kind of footprint is producing it: coverage, citations, or long-standing reputation.
- Watch for displacement, not only decline. A falling share of voice with steady Visibility means a competitor is being added to answers, not that you're being dropped from them. The remedies differ.
- Look at their sources. Sources shows the sites AI cites in your category. A site that's cited often and rarely names you is usually naming someone else.
- Don't over-read one check. Single-check differences are noise; patterns that hold across checks are signal.
Alerts on competitors
Alerts can tell you when a rival passes you, when share of voice moves, and when you drop out of an engine or are named by one for the first time. Each can be switched on or off per brand.
Choosing whom to track
Track the brands buyers actually compare you with, which isn't always your strategic archrival. Public profiles and the answers in your first checks show who engines name alongside you. If a brand you've never thought of as a competitor keeps appearing in your answers, that's the market telling you your set is wrong. To change the set, open Edit competitors at the top of the Competitors section, or the Tracking tab of Brand Settings. Each competitor has a name and a website, which is where its logo comes from. Anyone who can edit the brand can add, remove or rename competitors, up to your plan's limit.
What happens when you save:
- Past share of voice is recalculated with the new list, and with it the Prominence part of the Flare score and the Competitors section, for earlier dates too. Visibility, position, sentiment and accuracy don't change.
- A new competitor is fully counted from the next check. Answers already collected weren't checked for it, so until then its numbers are missing or incomplete, and the Competitors section says so.
- A removed competitor drops out of the history. Its mentions are kept, so adding it back brings them back.
- Renaming starts a competitor's history again, unless you only add detail in brackets, such as FAB (First Abu Dhabi Bank).
- No share alerts are sent across the change. The first check after it has nothing like-for-like to compare with, so alerts about share of voice and rivals passing you resume from the check after.
Where to go deeper
The method behind measuring rivals on one yardstick is in how do I track competitors in AI answers?, and Explore shows the Flare score of every brand in your category before you commit to tracking anyone.