Company · How we work
Map, automate,hand over, keep running.
Every engagement runs the same four steps, whether it is one workflow or a whole operation. Here is what happens in each, who is in the room, and what you own at the end.
The shape
One workflow at a time, end to end.
We do not start with a platform. We start with one workflow that costs your team time today, follow it from the moment it begins to the moment it is finished, and automate the whole of it. A half-automated workflow still needs a person to watch it; a whole one does not.
The people who do the work are in the room from the first day. They know where the exceptions are, which fields are always wrong, and which step everyone quietly skips. The system we build has to survive contact with that, so it is built with them, not presented to them.
When the first workflow runs on its own, we pick the next. Each one is scoped, priced and measured on its own, so you can stop at any point with something that works.
The four steps
What happens in each.
Map, automate, hand over, keep running. The steps are the same for every engagement; the size of each depends on the work.
- 01
Map
We sit with the people who do the work and write down what actually happens: every system touched, every hand-off, every exception. The map is a document you keep, whether or not we build anything.
- 02
Automate
We build the models, agents and integrations that do the carrying, and test them against your real records with a golden set we agree on before we start. Nothing goes live until it passes.
- 03
Hand over
Code, weights, prompts, evaluations and runbooks are handed over whole, in your accounts. Your team is trained to run and change the system without us.
- 04
Keep running
We watch what we built, re-score it as your data drifts, and keep a person on call for the exceptions, for as long as you want us to.
What we hold to
Rules we do not bend.
- 01
The record is the truth
Every system we build reads from your record and writes back to it. No side spreadsheets, no shadow databases, no copy of your data that only we can see.
- 02
Agents act as someone
An agent runs with the permissions of the person it works for, no more. Every action it takes is attributed, timestamped and reversible.
- 03
Measured before it ships
We agree on how to tell whether the system is right before we build it. If there is no way to measure it, we say so and do not build it.
- 04
People keep the decisions
Systems prepare, draft, route and reconcile. Anything with a consequence for a person — a diagnosis, a credit decision, a dismissal — is decided by a person.
- 05
Your keys, your region
The work runs in your cloud account, your identity provider and the region you choose. We hold no standing access once it is handed over.
- 06
Everything is yours
What we build for you is yours to keep, change and take elsewhere. Hand-over is part of the engagement, not an add-on.
Details
On the spec sheet.
- Start
- A conversation with an engineer, then a mapping session with the people who do the work
- Team
- A small team of engineers who scope, build and hand over the same work; no separate sales layer
- Scope
- One workflow at a time, end to end, each scoped and priced on its own
- Cadence
- A working session every week; a demonstration against real records at each milestone
- Evaluation
- A golden set agreed before building; re-scored as the data drifts after launch
- Deliverables
- The workflow map, the system, its evaluations, runbooks and the training to run it
- Ownership
- Code, weights, prompts, evaluations and runbooks handed over whole, in your accounts
- After hand-over
- Optional operations: monitoring, re-scoring and a person on call for the exceptions
- Pricing
- Fixed scope, priced before we start; SoloOne per module, per seat, per month
Questions
Asked before we start.
How long does a first engagement take?
It depends on the workflow, and we will not quote a number before we have mapped it. What we can say is that the map itself is short, the first automated workflow is scoped to be measurable, and you see it run against your real records before it goes live.
Do we need to replace our systems?
No. Most engagements sit on top of the systems you already run and carry information between them. Where the existing systems are the problem, SoloOne is there, and moving to it is its own scoped engagement.
What happens if the system is wrong?
It is caught. Every automated action is logged against the record and can be reversed; anything the system is unsure of is handed to a person with the context attached. After launch we re-score the system against the golden set as your data changes, so drift is found before it costs you.
Can our own team take over after hand-over?
That is the point of hand-over. Your team gets the code, the evaluations and the runbooks, and the training to change the system without us. Ongoing operations are there if you want them, not a condition.
How is the work priced?
Services are fixed scope, priced before we start, one workflow at a time. SoloOne is priced per module, per seat, per month; the configurator on the pricing page shows the arithmetic.
Start
Tell us what your teamstill does by hand.
We map the workflow with you and say plainly what can run on its own, what should not, and what it takes.