Modules are taken individually, so a firm that never opens the operations module is not billed for it. Minimum term is six months.
Runs whether you use it or not
- AWS infrastructure — hosting, the Australian model gateway, backups, monitoring, recovery$12,000 / year
- Assurance & audits — controls, access and data handling, the evidence behind them, the annual review cycle$6,000 / year
- Data foundation — the warehouse, the models, the quality layer underneath everything$20,000 / year
- Distribution module$20,000 / year
- Operations module$10,000 / year
Moves with your firm
- Integration — small$1,200 / feed / year
- Integration — large$2,400 / feed / year
- Data volume — storage and processing$200 / GB / year
- AI usage — at AWS Bedrock’s published ratesat cost
- Users — the people with an account$200 / user / month
- Consulting — beyond the subscription$1,500 / day
What we don’t charge for
- No implementation fee — the build already exists, so there is nothing to bill for standing it up.
- No margin on model usage — AI is billed at AWS Bedrock’s published rates, at cost.
- No per-conversation metering — seats are priced per person. A bill that grows when adoption succeeds is a tax on the thing working.
- No hire — the administration sits in the platform fee, not in a job description.
What you will actually be billed for
Three lines account for nearly all of it — the platform fee, the seats and the integrations. Every one of them is a person or a piece of work. The model usage itself, the AI everyone is actually asking about, is the smallest line on the invoice.
If two of our clients compared invoices, the rates would match and the counts would not — and both would be able to see why.
How that compares
We set the same scope against a US vendor stack and against building it in-house — line by line, first year, 20 seats. The gap is $326,000 in year one, and almost none of it is the AI.