Leading design for care

Alongside ongoing async feedback between the team, I run regular design crits as structured, scheduled sessions - because a team doing regulated, high-stakes work needs feedback they can rely on turning up. We're an open, honest team; where banter runs freely, so does constructive feedback, so a split opinion on a design decision is rarely shied away from and instead often inspires lively discussion.

On the whole, design decisions are litigated against research, or already tested and established solutions in Pulse (design system). If pressing or genuinely clear-cut, I'll step in and make the necessary call.

Most of my time is currently invested in monitoring across product areas, setting direction and standards, and diving into the design weeds myself including at canvas level. Understanding what the team is working with first-hand is pivotal to meaningfully evolving the lens I work through.

Three things anchor that right now.

Care Contextual Patterns

An interface without Care Contextual Patterns, even if accessible, still allows someone to make the wrong call under pressure - a medication round, a safeguarding conversation, a crisis moment where the wrong hierarchy costs more than a usability score. I defined Care Contextual Patterns as its own discipline, sitting alongside accessibility rather than folded into it: cognitive load, crisis-mode flows, neurodivergent and language-diverse users, across both the professionals delivering care and the people calling on it.

Accessibility is the floor. Care Contextual Patterns are the guardrail.

Language & Voice

Every product communicates, whether anyone designed it to or not - word choice, tone, imagery, often decided by whoever happened to write the copy. Today I’m leading the initiative on how our platform speaks, end to end: this includes content standards, imagery rules, tone registers, right down to the patterns teams actually integrate throughout the platform. How you word a safeguarding notification or an end-of-life journey isn't about style, it’s about dignity and safety.

AI output standards

Today we’re already reflecting on what it means to navigate and test with AI prototypes. Once AI starts generating content a care professional reads directly, someone has to own what it can say and what it must never claim - there must be guardrails around clinical judgement it doesn't have. This is a specific, verbal version of ‘AI governance’: disclosure language, tone, and hard limits on machine-generated content in a clinical setting.

None of this happens passively. Today, I’m working to set these standards with the team, and to ensure diligence even in the face of pressure to ship fast. This is a core part of my role: defining what people design to, and being accountable when it's actually tested.

Design sets the bar everything else is built against. This is the difference between design as a service and design as a standard.

Care Contextual Patterns, Accessibility, Language & Voice, and AI output standards sit with design, to decide what "good" means - driven by research and testing, and never shaped in isolation. Alongside user dialogue, these standards are defined in partnership with Clinical Safety, engineering, and PMs who are dedicated to the problem spaces we serve.

Hiring here is close to constant, driven mostly by growth rather than one-off backfill. Since starting at Nourish in late 2024, I've reviewed roughly 470 applications (380 design, 90 UXR), conducted ~90 full interviews plus another 10 lightning rounds, and made more than 10 permanent hires. At that volume, my process has to hold up under repetition: the same structure, the same evaluation criteria, so that a yes means the same thing in January this year as it does next July.

Bringing people in is only half the story. The other half is keeping them empowered, fulfilled, and constantly progressing once they're here. Alongside regular 121s and team-wide touchpoints, I track capability against a multi-level competency framework built around what each role actually needs, and what design progression should look like for individuals and teams.

I'm building our AI playbook on five principles, namely:

1. AI output is a starting point for discussion, never a prescription.
2. AI assists discovery, it doesn't shortcut it.
3. AI integrates into a human-centred process, rather than replacing stages of it.
4. AI helps to prioritise learning. Time saved goes back into research and iteration, not just more output.
5. Human at the helm. Every AI-assisted output gets assessed for bias, user alignment, and strategic fit before it ships.

Rolling out AI tooling to the team wasn't a mandate but a series of design sprints and even ad-hoc experiments, testing Claude and Figma MCP (and initially a few other emerging AI tools) against real build problems, and reviewing our findings. Output was discussed openly in crits, the same way any other design decision gets litigated against what Pulse already holds as a benchmark.

This approach was deliberate. The team now integrates AI-assisted workflows into day-to-day delivery without friction or quietly lowering craft standards.

Alongside this, I’m architecting a ‘Clinical AI Guardrails’ framework within our content design strategy, defining governance for generative UI in high-stakes environments. This ensures any AI-assisted care or clinical flow is fitted with the necessary guardrails for clinical safety, while any ‘Prompt as Interface’ patterns maintain clinical intent at the core of the LLM interaction.

Governance and a strong design system foundation is what makes this fast and trustworthy, rather than fast and quietly inconsistent - or worse, unsafe.