Your prompt panel defines what Brandflare measures. Prompts are the AI era's keywords: the finite, repeatable questions your visibility is scored against. This page covers how the panel is built, what makes it trustworthy, and how to read the matrix that shows your results prompt by prompt, engine by engine.
What a prompt is here
A prompt is a question a real buyer in your category asks an AI assistant: "best X for Y", "X vs Y", "is X worth it", recommendation and comparison queries. Good panel prompts share two properties:
- They never name your brand. Asking "tell me about Acme" tests whether the engine knows you exist; asking "what's the best project tracker for small teams" tests whether the engine recommends you — which is the question that moves revenue.
- They're questions people actually ask. The panel should mirror real buying conversations, not marketing language.
How your panel is built
When your brand is onboarded, Brandflare generates its panel from your category and competitors. You review it as part of setup — the panel is your yardstick, so it's worth confirming the prompts cover the buying questions you care about, at your plan's panel size (15 prompts on Starter, 40 on Trial, Pro, and Agency — see plans and limits).
After that, the panel stays fixed. This is deliberate: prompt tracking only produces meaningful trends on a stable yardstick. If the questions change every week, a rising score could mean rising visibility or just easier questions, and you'd never know which. Individual prompts can be deactivated — an inactive prompt simply stops being swept — but the discipline of a mostly stable panel is what makes month-over-month movement attributable to the world changing rather than the ruler changing.
The matrix
The prompt matrix is the panel's results laid out as a grid: one row per prompt, one column per engine — ChatGPT, Gemini, Perplexity, and Claude. Each cell answers the atomic question: did this engine's answer to this prompt mention your brand?
Where you were mentioned, the cell also carries your rank — your position among the brands the answer named, as determined by the judge. Being the first brand recommended and being the seventh name in a comparison table are both mentions, but they are very different outcomes, and rank is how the matrix distinguishes them.
Rank movement arrows show change between sweeps: an up arrow means you moved earlier in the answer for that prompt-engine cell since the previous measurement, a down arrow means you slipped. Because engines are non-deterministic, treat a single arrow as a hint and a persistent direction across sweeps as a finding.
Every cell is backed by a stored answer. Click through and you can read the verbatim response that produced the result — which brands it named, in what order, in what framing. Nothing in the matrix is more than one click from its receipts.
Reading the matrix well
The matrix is where the headline AI Visibility Score stops being a single number and becomes a diagnosis. Patterns worth looking for:
- Empty rows — prompts where no engine mentions you. These are your losses, and your content roadmap: each one is a buyer question your material doesn't currently win.
- Empty columns — one engine that consistently omits you while others don't. Engines disagree because their training data and retrieval differ; a lagging engine tells you where your footprint is thin.
- Mentioned-but-last — rows where you appear with a consistently poor rank. You're in the consideration set but framed as an afterthought; these often respond faster to work than empty rows do.
- Volatile cells — flipping between mentioned and absent sweep to sweep. You're on the boundary of the engine's answer; marginal improvements in the sources it draws on can tip these your way.
How the matrix feeds your metrics
Each cell is one observation; your mention rate is the fraction of cells (per engine, per day) where you were mentioned, and the AI Visibility Score averages that across engines over the trailing window. Competitor mentions are extracted from the same answers, powering share of voice and head-to-head comparisons in competitor tracking. The cadence at which the matrix refreshes — daily or weekly — is your plan's sweep frequency, detailed in how sweeps work.
Panel size and plan
Starter's 15-prompt panel covers your core buying questions; the 40-prompt panels on higher tiers add breadth — more sub-niches, use cases, and comparison angles, and correspondingly more stable metrics, since every additional prompt is another observation averaging out engine noise. If your category is broad, panel size is usually the most consequential difference between tiers.