Nuffield Health – Designing AI-Assisted Healthcare Journeys (2026)

Exploring how conversational AI can improve patient decision-making, consultant discovery and digital healthcare experiences

Introduction

Nuffield Health launched its first patient-facing AI chatbot to support consultant discovery and booking. My role was to evaluate whether conversational AI genuinely improved the healthcare journey and identify where AI could create meaningful value for both patients and the business.

While the chatbot represented an important step towards AI-assisted healthcare, the wider opportunity was understanding whether conversational AI genuinely improved patient journeys, reduced friction and supported better healthcare decisions.

Rather than focusing solely on improving chatbot interactions, the project evolved into exploring how AI could become a guided healthcare experience, helping patients navigate complex treatment pathways, discover the right consultant and progress confidently towards booking.

Working closely with Product, Engineering, Digital teams and external AI partners, I helped shape the UX direction, conversational patterns, research strategy and future product opportunities.

The Challenge

The challenge wasn’t simply improving the chatbot interface.

The business wanted to understand whether AI could genuinely improve the healthcare journey while also supporting wider organisational goals.

These included:

  • Increasing digital self-service
  • Helping patients choose the right consultant faster
  • Reducing unnecessary contact centre enquiries
  • Increasing confidence throughout healthcare decision-making
  • Demonstrating measurable value from the AI pilot

Healthcare journeys are often emotionally complex. Patients may not know what treatment they need, which consultant is appropriate or even what questions to ask.

The opportunity was to determine where AI genuinely added value and where traditional UX patterns remained more effective.

Discovery

Before designing new concepts, I wanted to understand where users were struggling today and whether AI was solving the right problems.

The discovery phase included reviewing:

  • Existing chatbot journeys
  • Consultant discovery journeys
  • Treatment pages
  • Booking journeys
  • Hospital information architecture
  • Mobile experiences
  • Accessibility
  • Existing conversational patterns

Alongside stakeholder discussions, this helped identify opportunities where AI could simplify healthcare decisions rather than simply automate conversations.

My Role

As Senior UX/UI Designer, I partnered with Product, Engineering, Digital teams and external AI suppliers to evaluate the existing chatbot experience and explore future opportunities for AI-assisted healthcare journeys.

My responsibilities included:

  • UX strategy
  • Product thinking
  • Conversational UX
  • Journey mapping
  • Interaction design
  • HTML prototyping
  • Accessibility improvements
  • Moderated usability research
  • Insight synthesis
  • Future product recommendations

Hybrid Conversational UX

A key focus of the exploration was moving beyond traditional text-only chatbots towards a more structured conversational experience that combined the flexibility of AI with established UX patterns.

Rather than expecting users to type complex healthcare questions, the experience aimed to reduce cognitive effort by providing guided interactions that helped people discover information more quickly and confidently.

The concepts explored included:

  • Guided conversation buttons
  • Recommendation cards
  • Quick actions for common healthcare tasks
  • Structured navigation within conversations
  • Contextual suggestions based on user intent
  • Progressive disclosure to reveal information in manageable steps
  • Smart prompts to help users ask better questions
  • Personalised follow-up recommendations
  • Clear pathways to human support when required

The objective was to combine the speed and flexibility of conversational AI with the clarity and predictability of traditional user interfaces.

This hybrid approach helps users understand what the AI can do, reduces unnecessary typing, improves discoverability of services, and creates a more accessible experience for people who may be unfamiliar with conversational interfaces.

Why this approach?

Healthcare decisions often involve uncertainty. Users may not know the correct medical terminology or even what question they should ask.

By providing structured guidance alongside conversational AI, users receive support throughout the journey rather than facing an empty text box with little direction.

The exploration focused on helping users:

  • Find treatments more easily
  • Discover relevant consultants and services
  • Navigate hospital information faster
  • Receive personalised recommendations
  • Complete common tasks with fewer steps
  • Escalate to a human advisor whenever needed

Design Principles

The conversational patterns were designed around several core UX principles:

  • Guide rather than expect — reduce reliance on free-text input.
  • Show rather than tell — use cards, buttons and visual components wherever possible.
  • Reduce cognitive load — present only the most relevant options at each stage.
  • Keep users in control — allow them to change direction or browse independently at any point.
  • Build trust — make AI recommendations transparent and provide clear escalation routes.
  • Support accessibility — ensure interactions work equally well across desktop, tablet and mobile devices

Moderated User Testing

A moderated usability study has been designed to better understand behavioural patterns, trust, and confidence within AI-assisted healthcare journeys, helping inform future conversational UX improvements.

Key Research Insight

One of the biggest discoveries came through moderated usability testing.

Participants rarely wanted a traditional chatbot.

Instead, they wanted guidance.

Users consistently preferred experiences that helped them:

  • Understand their options
  • Find the right consultant
  • Receive personalised recommendations
  • Feel reassured before booking
  • Progress with confidence

This shifted the project from improving chatbot conversations towards designing guided healthcare journeys.

Strategic Pivot

The project evolved significantly as research findings emerged.

Originally the focus was:

Improve the chatbot experience

Research revealed a much bigger opportunity:

Design an AI-assisted healthcare guide

Instead of expecting users to know what to ask, AI could proactively support decision-making through guided interactions, recommendations and clearer pathways.

This became the foundation for future product exploration.

Design Principles

Research established four principles that shaped the future direction.

Guide, don’t search

Help patients recognise relevant healthcare options rather than expecting them to search through large amounts of information.

Reduce cognitive effort

Replace unnecessary typing with buttons, recommendations and structured guidance.

Build trust

Explain recommendations, surface consultant expertise and provide reassurance throughout the journey.

Keep humans available

AI should complement healthcare professionals, not replace them, with clear escalation whenever users need additional support.

VIEW HTML PROTOTYPE

UX Exploration

Rather than relying solely on free-text conversations, I explored a hybrid conversational experience combining AI with familiar interface patterns.

Concepts included:

  • Guided buttons
  • Recommendation cards
  • Suggested questions
  • Quick actions
  • Consultant recommendations
  • Treatment guidance
  • Contextual next steps
  • Human escalation
  • Mobile-first interaction patterns

The goal was to reduce friction while helping patients make confident healthcare decisions.

Moderated User Research

To validate the concepts, moderated usability testing was conducted with healthcare decision-makers across both desktop and mobile devices.

The research evaluated:

  • Current healthcare website
  • Existing AI chatbot
  • AI-assisted prototype

Research overview

  • 8 moderated usability sessions
  • 4 desktop participants
  • 4 mobile participants
  • 45–60 minute sessions
  • UK healthcare decision-makers

Participants completed realistic consultant discovery and booking tasks while researchers observed natural behaviour, confidence and decision-making.

Importantly, participants were not initially told the study focused on the chatbot, allowing genuine behaviour to emerge without influencing results.

Research Findings

The findings consistently showed that guided experiences outperformed traditional conversational interfaces.

Guidance beats conversation

Users preferred recommendations over typing questions.

Buttons reduce effort

Participants consistently favoured button-led interactions that reduced typing and cognitive effort.

Trust matters

Consultant profiles, recommendations and explanations increased confidence before booking.

Hybrid AI performs best

Combining conversational AI with structured interface components created the strongest overall experience.

Customer Voice

Several themes appeared consistently across every moderated session.

“I’d rather them just be buttons.”

“A mixture of buttons and typing works best.”

“I feel more confident after seeing the consultant’s profile.”

Participants weren’t looking for longer conversations.

They wanted reassurance, guidance and confidence.

Recommendations

The research identified several opportunities for future optimisation.

Immediate Improvements

  • Improve chatbot discoverability
  • Reduce typing
  • Introduce guided buttons
  • Surface consultant recommendations earlier
  • Explain AI recommendations
  • Improve accessibility
  • Simplify language

Future Opportunities

  • Consultant comparison
  • Suggested questions
  • AI conversation memory
  • Richer personalised recommendations
  • Improved handoff to human support
  • AI-assisted consultant discovery

Outcome

The project helped establish a broader vision for AI-assisted healthcare journeys.

Rather than viewing AI as a standalone chatbot, the work explored how conversational AI could become an integrated healthcare guide that supports consultant discovery, treatment navigation and informed decision-making.

The moderated research validated the strategic direction of combining conversational AI with structured guidance, recommendation cards and contextual interactions to reduce effort while increasing confidence.

The work also created a roadmap for future optimisation through continued user research, A/B testing and phased product improvements.

What I Learned

This project reinforced that successful AI experiences are not measured by chatbot engagement alone.

The real value comes from helping people make better decisions.

Patients consistently valued guidance, trust and reassurance more than open-ended conversations.

The biggest opportunity wasn’t building a smarter chatbot.

It was designing a smarter healthcare journey.

Final Reflection

This project changed the way I think about AI product design.

Rather than asking:

“How can we make users use AI?”

The more valuable question became:

“How can AI help users achieve their goals with less effort and greater confidence?”

That shift in thinking transformed the project from a chatbot optimisation exercise into the exploration of a broader AI-assisted healthcare experience, one that puts patient needs, business goals and evidence-based design at the centre of every decision.