Approach
How we run an engagement.
Five stages. Each one ends with something you receive and keep.
Delivery
Five stages, each with a deliverable
-
Discover
We walk the process with the people who run it, and count where the time and cost actually sit.
You get: Process map, ranked shortlist, written recommendation -
Prototype
A working version within days, run against real cases from your own history.
You get: Prototype, the cases it was tested on, what it handled badly -
Build and integrate
The production version, connected to your systems, in accounts registered to you from the start.
You get: Finished system in your accounts, with documentation -
Launch
A limited pilot, measured against the success criteria agreed in writing before we built anything.
You get: Pilot results, and a recommendation to widen, adjust or stop -
Operate and extend
We monitor it, report on how it performs, and extend it — or hand it fully to your team.
You get: Monitoring, performance reports, agreed response time
Principles
Four commitments, every engagement
Full ownership and handover
Accounts, keys and data stores are registered to you from day one, and documentation is written as we go. End the relationship tomorrow and nothing needs to move.
Governance from day one
Who sees what, where data sits, how long it stays, and where the system stops and asks a person. Decided at the prototype stage, while it's still cheap to change.
Measured delivery
Success is agreed in writing before we build, and reported against afterwards — including when it falls short. Uncomfortable for us, and the only honest test.
What we won't do
Recommend automating a process that doesn't pay back. Build something we can't hand over. Or claim a system is more autonomous than it is.
Practical questions
What working with us looks like
| How long does a first project take? | Two to three weeks for discovery and a prototype. Four to eight weeks for a first production build, depending on how many systems it touches. |
| What do you need from us? | Access to the process and the people who run it, sample data, and one person who can make decisions. The last one matters most. |
| Where does our data go? | Your own accounts, in US regions where the provider offers them. Any third-party service is named and agreed first. Sensitive records are built and tested on redacted copies. |
| What if it doesn't work? | The pilot exists to find that out cheaply. If the result doesn't justify going further we recommend stopping, and you keep everything built so far. |
| Can our own team maintain it? | Yes — that's the intended end state. Support is available, and nothing we build depends on it to keep running. |
Start with a consultation.
Bring the process that frustrates you most. You keep the written summary whether or not you hire us.