Your per-seat plan. Set the size of the team that will use Claude and the price per seat.
List the workflows you’ll run, pick how complex each one is, and how often you’ll run it — scale the frequency to your number of accounts and entities. Each run’s cost comes from the model and the complexity you choose (we estimate the tokens for you); switch the model to reprice.
Token estimates are based on a typical Numeric user; swap in your own figures with Edit tokens manually. A note on pricing: sending data in is cheap, but getting answers out costs more — output tokens are billed at roughly 5× the input rate. So a workflow that writes a lot (long flux narratives, detailed memos) costs more per run than one that mostly reads and returns a short answer.