Grouped by who usually asks. If yours is not here, write to us and we will answer it in a sentence rather than book a call.
No. The audit is the strategy. Two weeks watching real work beats a quarter of workshops, and you finish holding a costed build plan rather than a slide deck.
Below roughly thirty people the arithmetic rarely works, because the repetitive volume is not there yet. We will tell you that on the first call rather than sell you an audit.
Two hours a week from an operations lead, read access to the systems in scope, and permission to sit with the team while they work.
Through your accounts, under your credentials, with the run log in your infrastructure. We hold no copy of your customer data at any point.
Read by default, and write only for the specific objects its workflow touches. Your IT lead approves the scope before anything connects.
Disable the service account and the agent stops. Nothing else in your stack depends on us, so nothing else breaks.
It routes to a person. The handoff rule is written before the prompt, and anything the eval suite has not proven stays behind that gate until it does.
Whichever passes the evals for that workflow at the lowest cost per run. We change them when something better appears, and the suite is what makes that safe.
Yes, and some clients do. The prompts, evals, and connectors are yours. Managed operations exists because most teams would rather not staff it.
A $6,500 audit credited against your first build, agent development from $18k per workflow, and managed operations at $2,900 a month however many agents are running.
You do not pay for it. That has happened twice since 2021, both times because the team had already automated the obvious work themselves.
Model and inference costs under $400 a month are included. Above that we pass through at cost, itemised per workflow, with no markup.