Pricing
What VertixOS costs, and how we work it out.
It depends on four things, and the biggest by far is what you would let an agent touch. So we publish the bands we work in rather than a package table that would be wrong in both directions — and you can answer the biggest question here, before you give us anything.
The biggest mover
What would the agent actually be touching?
Pick the furthest one that applies. It sets the risk tier, and the tier decides how much a person stays in the loop.
Pick one and we will show you where it lands on the tier table, before you give us anything at all.
The bands
Roughly what each band buys.
Published so nobody has to guess whether they are in the right place. You pick yours in the qualification and the answer comes back scoped to it, which means the call is about the work rather than the money.
Under $5,000 a month
Below where we can do this well. Governed delivery carries real engineering and real operating cost, and we would rather say so here than discover it together on a third call.
$5,000 to $10,000 a month
Usually one workflow taken properly: measured first, built inside its approval rules, then run. A single system rather than several at once.
$10,000 to $30,000 a month
A full system live with a scorecard, plus us operating it: monitoring, the learning loop, policy tuning and a monthly governance report. The common shape for an operating company.
$50,000 a month and above
Several systems, a regulated or multi-entity rollout, or a programme with your own team embedded — with the deployment and evidence work that regulated buyers need alongside it.
The mechanism
What actually moves the number.
Four things, roughly in order of weight. This is why the same workflow is a fortnight of work in one business and a quarter in another — and why none of it is knowable from a web page, while all of it is knowable from five questions.
- 01
The risk tier of the actions involved. Anything that reaches a customer or moves money carries approval and evidence work that reading and drafting simply does not. It is the question at the top of this page, and it is most of the scoping.
- 02
How many systems the agents have to reach into. Every connector is a permission boundary as well as an integration, so the fifth one costs more than the first.
- 03
Whether the answers and the policies already exist. If your approved content and your rules are written down, we configure. If they are not, we help you write them, and that is real work rather than a footnote.
- 04
Whether you want us operating it afterwards. A system you run and a system we run are different commitments with different monthly shapes. You can change your mind later in either direction.
The shapes
Four ways an engagement runs.
Most start at the top and stop wherever the evidence says to stop. Nothing here obliges you to the next rung, and the measurement is worth having on its own.
- DiagnoseTwo to three weeks
The X-Ray
We read your own conversations and operations, privacy-first, and hand you the numbers: how long people wait, what never gets answered, what repeats, and exactly which share an agent could take — and which actions would still need a person.
- Build, fastOne week, on site or with your team
Governed Agent Bootcamp
One real workflow live in your own workspace, with its approval policy, its audit trail and a run-book — then a go or no-go you can defend to whoever asks.
- Build, fullSix to eight weeks
Governed Pilot
One complete system live on the operating system: agents, approval rules, your own branded console, a shadow run against your real traffic, and a scorecard you can put in front of a board.
- OperateOngoing, reviewed quarterly
Managed operations
We run the agents: monitoring, the learning loop, policy tuning, cost and model routing, and a monthly governance report with a named operator who answers the phone.
The qualification tells you which of these fits, and the reply tells you what it costs.
The model behind the number
What we will not do to you.
No per-seat pricing. People who only approve things are never a line item. Charging for them would penalise the exact behaviour we are selling.
No credits that expire. Unused allowance is not a product feature, it is a way of billing for nothing.
No overage multiplier. If you go past an agreed level, it is at the same rate, not a punitive one.
No governance upsell. Tiers, approvals and the audit trail are how the product works. They are not an enterprise add-on, and every engagement has them.
No charge for our own retries. If the platform has to re-run something because a check failed, that is ours to absorb.
No surprise at the end of the month. The usage that models and tools actually consume is shown to you rather than marked up behind an abstraction.
Straight answers
The questions this page usually raises.
Tell us what the agent would touch.
That one answer does most of the scoping, and you will have a figure back before the end of the day.
Five multiple-choice questions. A person replies within one business hour.
Autonomy where it is safe. A human where it is not.