Home / What we build

What we build

Useful software starts with a real task.

We design and develop the part of a system that helps a team understand its data and move the work forward. That may be a dashboard, a custom workflow or a controlled AI integration. Maps are one of our specialisms, not a requirement for every project.

A defined contribution

What we agree before building

User
Who makes the decision or performs the task
Input
Which data, API or existing system supplies the facts
Action
What a person or AI client may do next, with which permission
Proof
How the result will be tested and accepted
Where to start

Two ways into a project

Choose the route that matches what you already know. Both begin with a named task and an honest delivery boundary.

01 / Defined interface work

One screen or operator module

Use a published review or module scope when the user task, existing platform and data source are already clear. The package page shows deliverables, exclusions and price.

See defined packages
02 / Scoped software work

Dashboard, internal tool or AI integration

Send the workflow and systems involved. Nick reviews fit, then we agree the questions, first-release boundary and any discovery work before giving a build price.

Discuss the workflow
Capabilities

Three ways we make operational work easier

01 / Dashboards

See what needs attention

Role-appropriate views for status, trends, exceptions and the next action. We connect charts, tables, maps and detail views to supplied data and design for missing, late and conflicting information.

Working product: explore Road Safety with public data. Client work: SOTERIA analysis views and SWARCO monitoring.

02 / Custom software

Move a workflow into a usable tool

A bounded web module or internal application can replace a handoff between spreadsheets, inboxes and disconnected screens. We design the user journey and build the agreed interface, server-side logic and integrations with the system owner.

Relevant proof: operator workflows and our internal CRM workflow.

03 / AI integration

Give AI a small, testable job

Connect an AI client to approved tools or add an assistant around a defined task. We specify what it may read or change, when a person reviews the result and which examples must pass before use.

First-hand proof: our MCP-connected acquisition system. It is an internal build, not a client deployment.

Where GIS fits

Maps when location changes the decision

SWARCO and SOTERIA show the depth of our map and live-data work. We use GIS when place, movement or spatial comparison is essential to the user task. Otherwise, the right interface may be a table, timeline, form or dashboard.

Explore operator interfaces ↗

How AI fits

Automation with an owner

AI can help draft, retrieve or classify information, but its outputs need a defined source, permitted action and review route. We keep the surrounding software and human workflow visible so the feature can be evaluated in use.

See the internal AI-connected workflow ↗

Delivery

From a narrow first release to a working handover

01

Frame the task

One user, one decision and the current friction.

02

Inspect the inputs

Representative data, API access, roles and constraints.

03

Build in context

Working screens and integrations reviewed with your team.

04

Test and hand over

Acceptance checks, known limits, documentation and ownership.

Published prices apply to the specific interface packages shown there. Wider custom development, back-end work and AI integrations are quoted against the actual system and scope.

What would make the work easier?

Send one example of the decision, the software involved and the outcome you need. We can suggest a first slice that can be built and checked.