How we work

One process, scoped and measured before anything is built

We begin with a single high-impact process and document everything before a line is written: the scope, the permissions, the escalation owner, and the success measures. Expansion follows results rather than preceding them.

Last updated: 12 August 2026
01

Engagement models

Engagements take one of three shapes, chosen on process maturity and whether an in-house technical team exists.

Scoped pilotA single process, a fixed scope, and success measures agreed in writing before the build begins. Suited to a first engagement.
Operations retainerA portfolio of agents monitored, tuned and extended each month, with a review pack and a named account owner. Suited to established programmes.
Build & handoverWe build, document, train your team, and transfer ownership. You operate the system from launch. Suited to in-house IT teams.
02

Delivery sequence

Four steps, each with a documented output before the next begins.

  • Step 01 — Process audit: we work alongside the team performing the task and document how it operates in practice, including the exceptions and the points at which work waits
  • Step 02 — Design & permissions: scope, system permissions, escalation owners and success measures are documented and signed off before any build begins
  • Step 03 — Build & test: the agent is built against live systems and tested on historical cases, so its accuracy is known before it engages with any customer
  • Step 04 — Launch & measure: launch takes place under supervision, with every action reviewable, and scope extends only once the measures are met
03

What governs the commitment

Response times, uptime, accuracy thresholds and escalation obligations are set out in our published Service Level Agreement and Service Standards.