Priced by the question, not the seat.

Why not charge per person, like everything else?

Because most of the people who need an answer are not the people asking. A regional manager who reads the weekly answer costs us nothing to serve — the work was done when somebody else asked. Charging for them would price the business out of sharing what it learns.

So a seat is somebody who may spend, and everybody else reads for free, on every tier. The product already draws that line internally: asking costs money and reading does not.

What happens when we run out of credits?

Asking stops until the first of the month, and the app says so plainly, naming the date. Everything already answered stays readable. It never keeps going and bills you afterwards — a charge nobody agreed to in advance is worth less than the renewal it costs.

Can we buy more part way through a month?

Yes — tell us and we move you up, immediately and prorated. If a month was unusual rather than a new normal, we would rather add credits than move you onto a tier you do not need.

Is there a free tier?

No. Every question runs real work against real systems and costs us real money, so a free tier would be us paying for strangers' questions. The Pilot tier exists instead: a quarter, one department, a real question, at a price that makes it a decision rather than a procurement exercise.

Do prices move with model costs?

What the models cost is our problem, not a line on your invoice. You buy credits at an agreed price and we manage the rest — including making the pipeline cheaper, which is work we should be doing anyway.

What does onboarding involve?

Connecting one or two systems, confirming how your business counts — fiscal year, currency, who is asking — and agreeing what each job title may read. On Team and above we do the first two connectors with you. Most customers are asking real questions within a week of the credentials arriving.