How it will work
What using it will look like
Seenby is not open yet, so these steps are the order it is planned to work in, not a sequence to start here.
Describe the brand and name the rivals
Enter your brand name, a short description, and up to three competitor names. Seenby turns that into 50 to 100 category prompt templates, with a curated fallback set if generation comes up short, and shows you the list before anything runs.
The week's cycle runs your prompts
The scheduled query engine sends the prompt set to OpenAI and Perplexity through a provider adapter, isolating per-prompt failures and continuing on the remaining provider if one is unavailable.
Responses are scrubbed, then parsed
A PII-scrubbing step runs in the storage pipeline before any response becomes queryable, and every stored response carries a scrubbed flag and the model version that produced it. Then mentions, sentiment and ranking position are extracted.
The score is computed with its caveat attached
Share of AI Voice is computed once per brand per week and stored together with the methodology caveat text, so every place the number appears can show what it does and does not support.
You read the digest, then the detail
The weekly email brings top insights and drift alerts. The dashboard holds the 30-day trend, the sentiment breakdown, the competitor comparison, and the prompts where nobody named you.
For what it covers, see the pages this site publishes.
A prompt set written for your category
Give us your brand name, a short description, and up to three competitor names.
Mentions, sentiment and position, extracted the same way each week
Every response is parsed for brand and competitor mentions, classified positive, neutral, negative or absent, and given a ranking position within the answer. Consistent parsing is what makes week-over-week comparison worth reading.
Questions
Answers before you ask
Which AI surfaces does Seenby query today?
Two at launch: OpenAI and Perplexity. They sit behind a provider adapter interface so a cycle degrades gracefully if one is disabled. Gemini and Claude are not included in the current scope.
How reliable is a Share of AI Voice score?
It is a repeated measurement of a non-deterministic system. The same prompt can return different answers on different days, and provider model updates can move a score without anything changing in your market. Seenby stores the model version behind every response and tags drift where it applies, and the methodology caveat is attached to every computed score rather than kept in a footnote.
How often does it run?
Weekly. The query cycle runs your prompt set, the parser scores the responses, and the digest summarises the week. Real-time alerting outside the weekly cycle is not in scope.
Where do the prompts come from?
Onboarding generates 50 to 100 prompt templates from your brand name, description and competitor names, with a curated fallback template set if generation fails or returns too few. A custom prompt builder is not part of the current scope.
What happens to personal names that appear in AI responses?
A PII-scrubbing step runs in the response-storage pipeline before any response is queryable, and every stored response carries a scrubbed flag. The privacy page documents the basis for processing generated content that references real people.
Does Seenby promise to improve how AI answers describe my brand?
No. Seenby measures and reports. What moves a generated answer is outside any tool's control, and we will not claim otherwise. The value is knowing your position, your competitors' positions, and which prompts leave you out.
Can I benchmark against competitors I choose?
Yes. You name up to three competitors during onboarding, and the parser extracts their mentions from the same responses as yours, which is what makes the comparison a like-for-like reading rather than two separate studies.
What does joining the waitlist involve?
An email address. We will write when access opens, with what is included and what it costs, and nothing in between beyond occasional notes on the category.
Join the waitlist
Leave an email and we will write when access opens, with the surfaces covered, the caveats that come with the score, and the price. No sequence, no drip.
One email when it launches, and one now to confirm your address. Unsubscribe any time. Privacy.