"Are you mentioned by ChatGPT" — yes or no. That's where we started too, and we kept running into the same problem: the yes/no was hiding the actual story. So we built four things to go past the single snapshot.
Citation Stability — Getting cited once? Could be luck. We started tracking whether it holds up across repeated checks, so you actually know if a win is real or a one-time fluke.
AI Engine Agree/Disagree — We ask Gemini, ChatGPT, Perplexity, and Claude the exact same question. Turns out they don't always agree with each other. We now show you exactly where they split, instead of collapsing that into one flat answer.
Mention & Citation Consistency — One scan is a snapshot, not a track record. We track how reliably each question gets answered the same way over time, so you know which numbers to actually trust and which ones to take with a grain of salt.
Retrieval Readiness — This one we're proudest of. For every question you're missing, we show you the page that got cited instead of yours, score both against the question, and surface real, checkable signals — page structure, whether that domain shows up elsewhere in your results — that help explain the gap. And we're upfront about what the score actually means: it's the closest-matching passage we could find, not a verified or complete answer to the question. A low score usually just means the page doesn't address that question yet — not that we missed something better.
None of this came from a whiteboard. Every one of these checks got built off a real scan, a real false positive, or a real "wait, why did that happen" moment — and we tested each one against real data before shipping it.
Curious where your own brand actually stands with AI? Try it free at https://www.webscore.dev/