A clear scope and acceptance criteria
A shared definition of the workflow, users, requirements and what a successful delivery means.
Let’s talk
From a defined problem to a working system
Custom AI systems, software and integrations built around the work you need to do. From the first clear brief through delivery and handover.
Discuss a bespoke buildIndependent advice. Direct access to the people leading the work.
Alongside our own ventures CNTXT_, Slypstream and SILVA, our client portfolio includes bespoke builds CTLG_ and TRMNL_.
Delaware / Cape TownA useful custom system has to fit into a real environment. People already have ways of working. Data lives in different places. Existing platforms have constraints. Someone has to support the result after launch.
I start by understanding that environment and agreeing the outcome. What should the system do? Who will use it? What must it connect to? How will we decide that it is ready?
From there, I design the smallest coherent solution and deliver it in reviewable stages. You can see the work, test the assumptions and make decisions before uncertainty turns into expensive rework.
For AI builds, the surrounding engineering matters as much as the model: access controls, retrieval, evaluation, exception handling, human review and visibility into how the system is behaving.
The engagement includes an agreed route to handover. Ownership, documentation, training, support and maintenance are discussed in the scope, so the operational responsibilities are clear.
Matt Watson
A clear scope
The details are agreed around your business.
These are the foundations of the engagement.
A shared definition of the workflow, users, requirements and what a successful delivery means.
The data, integrations, permissions and operating choices needed to work in your environment.
Implementation with agreed checkpoints, testing and feedback from the people who will use it.
Documentation and an agreed plan for operation, ownership and any ongoing support.
Have a challenge in mind?
Discuss a bespoke build ↗AI tools & integrations
Connect the AI tools your team uses to the systems your business runs on. We scope the connections, permissions and human review around the work you need to get done.




How we get started
We establish fit before proposing an engagement. Scope, fees and responsibilities are agreed before work starts.
Map the workflow, constraints, dependencies and acceptance criteria.
Deliver in stages, test assumptions and keep the business involved.
Validate the agreed requirements and establish how the system will be operated.
Before we talk
Rather ask than read?
No. Some problems are better solved with conventional software or an integration. We choose the approach around the workflow and the outcome.
Yes. Integrations are a core part of the work. We first establish the available APIs, access, data quality and constraints before agreeing a delivery scope.
Ownership, licensing and third-party components are set out in the engagement agreement before work starts. We discuss any requirements you have during scoping.
A credible estimate depends on the scope, integrations and unknowns. Share the outcome you need and the constraints you know; we will establish what is needed to prepare a proposal.
Start a conversation
Start with the problem, the people who need the system, and the tools it has to fit around.

What happens next
A few lines is plenty. No brief or deck needed.
Around 30 minutes with the person who would lead the work.
A proposal, a plan, or a straight answer if we’re not the right fit.