Skip to content

Method

How we work

The same four steps every time: Understand, Shape, Enable, Improve. The point is to reduce the number of surprises for everyone involved.

Understand. Shape. Enable. Improve.

  1. Understand

    Map the workflow, stakeholders, friction, cost, constraints, and what success would actually look like.

  2. Shape

    Redesign the process, then decide where automation, AI, integrations, or internal tools belong, and where they do not.

  3. Enable

    Implement, document, train, and preserve appropriate human control over consequential decisions.

  4. Improve

    Measure adoption and outcomes, maintain what was built, and expand only where the economics justify it.

Method and engagements

Which engagement belongs to which step.

The framework is not an abstraction laid over the commercial model. Each step is where specific engagements do their work.

Start with the work

Four questions before anyone builds.

A useful AI project begins with an operating problem that can be described, owned, and measured. These are the questions we use to find a sensible first step.

  1. 01

    Where does the work stall?

    Trace one request from arrival to completion. Mark the handoffs, duplicate entry, waiting time, and exceptions that people handle manually.

  2. 02

    Who owns the result?

    Name the process owner, the people who use the workflow, and the person who can approve changes when a decision carries risk.

  3. 03

    What is the baseline?

    Capture today's volume, cycle time, rework, errors, or missed follow-up. Choose a measure the team can review after the change.

  4. 04

    What can safely change?

    Identify the systems and data involved, access limits, exception paths, human review points, and what should remain a human decision.

The free 20-minute Fit Conversation checks whether the problem and working relationship fit. Detailed mapping, financial modeling, and an implementation plan belong in a paid AI Opportunity Map when the situation calls for one.

Choose the right first engagement

The starting point depends on what you already know.

There is no reason to buy a larger build before the workflow, ownership, and risk are clear. We select the smallest engagement that can produce a useful decision or result.

  • The problem is clear and bounded

    Build with You or a focused Build for You scope

    A named owner, known systems, a manageable workflow, and clear acceptance criteria may support direct scoping or a Collaborative Build Block.

    Explore this path

  • The priority or economics are unclear

    AI Opportunity Map

    We map the current process, compare opportunities, examine risk and payback, and recommend a responsible first implementation.

    Explore this path

  • The team needs capability first

    Teach You to Build

    A workshop or enablement sprint can give the team practice, approved-use boundaries, and a practical way to apply AI to its own work.

    Explore this path

What finished work includes

A working system needs an owner and a way to check it.

For implementation work, we agree on the boundary before building and review the result against it before handoff.

Defined

Scope, owner, inputs, outputs, approvals, exceptions, and acceptance criteria are explicit.

Tested

The intended path and important failure paths are checked with the people who run the work.

Transferred

The client receives the agreed documentation, account access, operating guidance, and training.

Reviewed

The baseline and post-change measure are compared at the agreed checkpoint; claims follow the evidence.

How an engagement starts

A short conversation first, then the smallest thing that answers the question.

It is free, and it is short. We are working out whether what you have is something we can help with, and which starting point fits. It is not a free assessment: the analysis, the mapping, and the recommendation are the work an AI Opportunity Map exists to do.

  1. Step 1

    Fit Conversation

    Free and short. We work out whether this is something we can help with, and which starting point fits.

  2. Step 2

    AI Opportunity Map

    When the answer is not already obvious. Paid discovery that decides what to build, and what not to.

  3. Step 3

    Build for You, Build with You, or Teach You to Build

    The named engagement that matches how much of the implementation you want to carry yourselves.

  4. Step 4

    Measurement and handoff

    Documentation, training, and a checkpoint once the new way of working has carried real work.

  5. Step 5

    What comes next

    An optional Care Plan, the next workflow, deeper enablement, or a larger implementation, only where the economics justify it.

You can start directly at a workshop or a build block when the requirement is already clear and low-risk. Ambiguous, material, or multi-system work goes back through the Map before fixed implementation pricing.

What this page covers

  • Discovery and requirements

    Map the workflow with the people who do it. Identify stakeholders, friction, cost, constraints, risk, and what success would actually look like.

  • Architecture and build

    Redesign the process first, then decide what becomes deterministic automation, what becomes AI-assisted, what gets integrated, and what stays human. Prototype, build, test.

  • Deployment and adoption

    Implement, document, and train. Preserve human approval where consequences matter. A system is not finished when the software works. It is finished when people can operate it.

  • Measurement and ownership

    Measure adoption and outcomes against the baseline. Hand over accounts, documentation, and the ability to change what was built. Expand only where the economics justify it.

Next step

The fastest way to understand the method is to watch it run on one of your own workflows.

Show Us How You Work