Available now for senior UX/UI, product and service design roles Get in touch →

ADA · Adaptive Design Approach

Shaped by experience.
Sharpened by AI.

After 25+ years across consultancies, corporates and start-ups, I’ve learned that no two engagements are the same. ADA is how I assess each one, shape how design is delivered, and decide where AI adds real value — with every output still governed by people.

See the delivery profile

Why I built it

A design process shouldn’t dictate how we work. The engagement should.

Three factors have shaped every engagement I’ve worked on, from regulated banking journeys to a start-up building its first product. I call them the PPE model.

P

Project

Different objectives, timelines, resources and delivery challenges.

What I learned Adapt the design process, team structure and priorities to meet delivery needs.

P

People

Different personalities, capabilities and motivations need adaptable leadership and collaboration.

What I learned Build trust, understand individual strengths and empower people to take ownership.

E

Environment

Every organisation has its own culture, processes, technology and ways of working.

What I learned Integrate into existing environments, challenge inefficiencies and introduce improvements without disruption.

How AI moulded it

Lessons from designing with AI, project by project.

Zentility · 2017 – 2022

Designing AI into the product

As Founding Product Designer I designed portfolio and customer dashboards with AI-driven projections of energy spend and usage, and industry-first pricing reports with AI market analysis. It taught me that AI output only earns trust when the design explains it.

Fidelity International · 2022 – 2023

Designing within a platform decision

I revised the risk platform prototype three times — not only from testing feedback, but to align with the technology direction once ServiceNow was selected as vendor. The technology setup shapes what design can deliver, so it has to be assessed up front.

KHDA Dubai · 2025 – 2026

Prototyping AI-assisted guidance

I designed and prototyped AI-assisted learner guidance, using AI in my own prototyping workflow, and tested it over four rounds with educators and school administrators (6–8 people per round). Development was paused for budget reasons, but the testing showed where AI guidance helped and where people needed control.

This portfolio · 2026

AI design-to-code, hands-on

I rebuilt this site with Claude Code: I set the direction, content and design decisions, AI generated and iterated the code, and I reviewed every change in the browser before it went live. It’s stage 07 of the delivery profile below, in practice.

The approach

Assess. Adapt. Empower. Then reassess.

  1. 01

    Assess

    PPE model assessment. Project requirements and constraints; people, skills and stakeholder alignment; environment, tools and governance — including how ready the technology setup is for AI-driven design.

  2. 02

    Adapt

    Design delivery profile. Configure roles and responsibilities, adjust processes and collaboration, set review and decision checkpoints — and decide which stages are human-led, AI-assisted or AI-generated.

  3. 03

    Empower

    Adaptive design delivery. Give designers clarity and autonomy, remove blockers and reduce rework, and train and strengthen the design team.

Continuous assessment & adaptation — reassess at milestones, reviews and significant project changes.

An example delivery profile

Where people lead, where AI assists, and how it stays governed.

This is the profile I’d set for an AI-ready engagement. Stages marked In practice are ones I already work this way; Proposed stages are where ADA points next, adopted only when the assessment says the environment is ready.

Ownership keyHumanHuman + AIAIColour shows the primary delivery mode; all outputs remain subject to human governance.
01

Project approved

  • C-suite strategy
  • Market and sales pressures
  • Product roadmap
  • Operational pressures
In practice
02

Project scope

  • Vision and scope (typically a deck)
In practice
03A

Org documents

  • Infrastructure architecture
  • Business, compliance and regulatory rules
  • Operations

Confluence

In practice
03B

BA documents

  • Baseline process flows
  • Project rules and expectations

Confluence / Word

In practice
03C

Stories & criteria

  • Epics and stories defined

Jira

In practice
03D

Service designs

  • Detailed flows

FigJam

In practice

Claude Code / Codex — AI working alongside the team from here

04Human + AI

AI functional spec

  • Decision tables and rules
  • Conditional journey logic
  • Validation and error states
  • Requirements traceability
  • Human review and approval
Proposed
05AHuman + AI

Design library

  • Figma tokens, components and variants
  • Interaction states
  • Accessibility patterns
  • Responsive layouts
In practice
⇄
05BHuman + AI

Code library

  • React / React Native components
  • Tokens and component APIs
  • Accessibility and Storybook
  • Automated component tests
Proposed

06 — UX/UI design journeys Three alternative creation approaches — all require design review

06AHuman

Human-led

  • Designer creates and refines Figma journeys
  • Using approved rules and components
In practice
06BHuman + AI

Hybrid, AI-assisted

  • AI proposes journeys; designer refines
  • Interactions, accessibility and usability
In practice
06CAI

AI-generated

  • AI creates candidate journeys from specs
  • Designer reviews and approves before use
Proposed
07Human + AI

Coded prototypes

  • Functional React journeys
  • Conditional forms and navigation
  • Validation and mock services
  • Using approved design and rules

Claude Code (in practice) · Codex (proposed)

Partly in practice
08Human + AI

Testing & validation

  • User testing and analytics
  • Accessibility and design QA
  • Business-rule validation
  • Automated journey tests
  • Change requests and defects
Partly in practice
09Human + AI

Engineering

  • APIs and integration
  • Security and performance
  • Production implementation
  • QA / UAT, deployment and monitoring
Proposed
Continuous change management Risk, compliance, budget, user research and technical changes → impact analysis → stakeholder approval → updated requirements, Figma, code and tests.

Principles

What stays constant.

01

People govern

AI proposes; designers, product owners and engineers decide. Nothing ships without human review.

02

Fit the environment

Adopt AI where the technology, data and governance allow it — not because it’s available.

03

One source of truth

Rules, design library and code library stay in sync, so changes flow through requirements, Figma, code and tests together.

04

Test with real users

Prototypes — human-made or AI-assisted — are validated with the people who will use them.

Let's talk

Have a complex journey that needs simplifying?

I'm open to senior UX/UI, product and service design roles and contracts. Drop me a note and I'll get back to you quickly.

LinkedIn ↗ X / Twitter ↗

Please enter your details.