We work with you, not for you.
The people who live with the system sit in the reviews. Nothing is designed behind a curtain.
Methodology
Six stages · Five principles
Create for humans.
Our method in one line: understand the operation, design for the person doing the work, and prove it on real data before anything is built — with AI speeding up every step in between.
- 01Discover
- 02Define
- 03Design & build
- 04Review
- 05Deploy & train
- 06Support
Discover → Support
One loop, six stages
01 — Principles
Five rules we don’t bend.
Operations before interfaces.
We map who orders, who approves and what the data holds before we draw a single screen.
Real data, early.
Prototypes run on your products, references and prices — not lorem ipsum — so feedback is about the work, not the placeholder.
AI is the accelerator, not the product.
It compresses research synthesis, prototyping and content production. Judgement stays with people.
Accessible by default.
We design to WCAG 2.1 AA from the first screen, not as a fix at the end.
02 — The method
What happens at each stage, and what you hold at the end of it.
Stage 1: Discover
Discovery workshop, site visit and process audit. We learn the operation the way your team runs it, not the way the org chart describes it.
Methods
Stakeholder interviews · process audit · journey mapping · data review
AI speeds up
Research synthesis — interviews and logs clustered into themes
You hold
Audit report and priority list
Stage 2: Define
Concept and direction. We agree what success looks like, how information is organised, and which journeys matter most.
Methods
Information architecture · user flows · success measures · direction boards
AI speeds up
Product discovery — more directions explored before one is chosen
You hold
Signed-off direction and site map
Stage 3: Design & build
High-fidelity design, development and automation, in parallel where possible. You see working screens, not static pictures.
Methods
Design system · clickable prototype · front-end build · workflow automation
AI speeds up
High-fidelity prototyping, content and product imagery
You hold
Prototype, Figma library, working build
Stage 4: Review
Two structured review rounds are included. Feedback is collected in one place, answered in writing, and closed before we move on.
Methods
Review sessions · usability tests with real users · decisions log
AI speeds up
Test analysis — sessions transcribed and issues grouped
You hold
Approved design and change log
Stage 5: Deploy & train
Testing, deployment and training. Quality checks run before every release, and your team learns the system before it goes live.
Methods
QA testing · accessibility checks · go-live plan · team training
AI speeds up
Training notes and help content drafted from the build
You hold
A live system and a trained team
Stage 6: Handover & support
Documentation, ongoing support and a look at what changed. What we learn here becomes the next round of work.
Methods
Handover pack · support window · usage review
AI speeds up
Reporting and dashboard automation
You hold
Documentation you can own
03 — Three ways in
Same method, sized to the problem.
3–5 days
UX/UI sprint
Stages 01–03, compressed: a discovery workshop, UX recommendations, an interactive prototype and a design direction.
6–16 weeks
Full product
All six stages, end to end, for a portal, back-office or configurator.
Automation track
Audit, optimise, automate, report
The same method applied to a process instead of a screen.
See AI automation04 — Quality gates
Nothing moves on until it passes.
- Direction signed off before high-fidelity design starts
- Key journeys tested with real users before build
- Contrast and keyboard checks against WCAG 2.1 AA
- A dedicated QA pass before every release
- Handover documentation your team can work from
05 — Toolkit
What we work in.
- Design & review
- Figma · Adobe Creative Suite
- Build
- React · Next.js · WordPress
- Automation
- n8n · Make · WhatsApp Business Platform
- Data & reporting
- Airtable · SQL · Metabase · Looker Studio