Skip to content
AI Creative home

Process

How a business problem becomes a working system

Five stages from proposal to delivery. The point of the sequence is to get something real in front of real users early, then build outward from what proves out.

The sequence

Proposal, discovery, prototype, implementation, delivery

Each stage produces something you can review, and each one carries the decisions of the last.

  1. 1

    Proposal

    The problem, scope, sequence and acceptance criteria, agreed before work is committed.

  2. 2

    Discovery

    Current workflow, systems, data and roles confirmed against what the build has to handle.

  3. 3

    Prototype

    A working version of the critical path, validated with real users before scale.

  4. 4

    Implementation

    Built in reviewable increments with integrations, permissions and reporting in place.

  5. 5

    Delivery

    Release, migration, training and handover, with support and a clear owner for what's next.

Between the stages: digital contracting, immersion with your team, continuous visibility into the build, and end-to-end human testing before release.

Inside each stage

What actually happens

The detail behind each stage, and what you should expect to see from us while it runs.

  1. 01

    Proposal

    The problem, scope, sequence and acceptance criteria, agreed before work is committed.

    • A written statement of the problem and the operational outcome
    • Scope, sequence and what is deliberately out of scope
    • Commercial terms, milestones and acceptance criteria
  2. 02

    Discovery

    Current workflow, systems, data and roles confirmed against what the build has to handle.

    • Time with the people who actually run the workflow
    • Systems, data, integrations and permissions mapped
    • Edge cases and exceptions surfaced before they become rework
  3. 03

    Prototype

    A working version of the critical path, validated with real users before scale.

    • Interactive screens and workflow, not static wireframes
    • Reviewed with the people who will use it daily
    • Assumptions corrected while change is still inexpensive
  4. 04

    Implementation

    Built in reviewable increments with integrations, permissions and reporting in place.

    • Increments you can see and use, not percentage-complete reports
    • Integrations, roles, auditability and reporting built in
    • Automated tests plus real-world human testing on every increment
  5. 05

    Delivery

    Release, migration, training and handover, with support and a clear owner for what's next.

    • Data migration and cutover planned with the operation
    • Training and documentation for each role
    • Support and an agreed owner for the next round of improvements

How we work

The rules the process is built on

  • Working software over documents

    Progress is demonstrated in the product. Feedback on something real is more specific and arrives sooner.

  • Senior people stay on the work

    You are not buying expertise in the pitch and receiving account management in delivery.

  • Scope decided against the operation

    Every feature has to earn its place against the workflow, the users and the economics.

  • Acceptance defined up front

    Each milestone has criteria agreed before it starts, so 'done' is not a negotiation at the end.

  • Integration treated as first-class

    Ownership of records, validation, retries and exceptions are designed, not discovered at launch.

  • Built for what happens after launch

    Systems are designed to be changed, because the operation keeps changing after go-live.

Before any of this

If the assessment says a purpose-built system is the wrong answer, we say so. Buying, extending, integrating or automating is often the better decision, and it is cheaper to reach that conclusion in the proposal stage than after a build.

Start with the proposal stage

A short conversation about the workflow, what it costs you today and what a first milestone would need to prove.