About
who we are
A specialist practice, not a general AI consultancy
We do one thing: put Claude to work inside IBM Business Automation Workflow — as assistant tasks that run as steps in a process, and coach views that appear in the screens your people already use.
Enterprise AI usually sits beside the work. Someone leaves the system of record, pastes context into a chat window, gets an answer and carries it back by hand. It demonstrates well and changes nothing about cycle time. We build the other thing — components that live inside the process, inheriting the case, the assignment, the escalation and the audit trail.
That focus is deliberate. IBM BAW is a specific platform with specific mechanisms — external implementations, message events, coach views, team filter services — and using them well is not something a generalist picks up on the way past. We publish the architecture we build to, so you can judge it before you talk to us.
why you should work with us
One specialism:
IBM BAW and Claude. We are not a general AI shop, and we will tell you when your problem is not ours.
The model never routes:
Claude classifies, extracts and drafts. Where a case goes next is decided by deterministic rules over an immutable published version — so it is still explainable a year later.
Your data stays yours:
For configuration work Claude receives a description of your data structures, never the rows. Not sampled, not anonymised — none.
Flows are data, not deployments:
Changing a process should not require a release. We build so that it does not.
It still works without the model:
Every AI feature we build has a defined behaviour when the model is unreachable. Usually that is: route to a human and keep going.
Everything we claim is written down:
Our architecture whitepapers set out the safety model in full, including where it does not apply. Read them before you call us.
Technologies we work with
