top of page
2024
How I Used AI to Create an AI for Better Help
Designing Echo, an AI assistant inside SocialHeads to transform social care through intelligent support.
Tools
Figma, Miro, GPT-4
My Role
Lead Designer
Team
4 members
Duration
8 weeks
Project Overview
Focus Areas
Research, UX/UI, AI Integration, Prototyping
Timeline
6 months
My Role
Lead Product Designer
SocialHeads is a secure digital platform designed specifically for social care—a safe alternative to everyday messaging apps that meets safeguarding standards while remaining approachable for young people.
Echo is the AI-powered assistant at its heart, designed to fill critical gaps in care while supporting overworked social workers with intelligent tools that save time and highlight risks.
Key Insights
The System
•
Warning signs get buried in paperwork
•
Risk assessments happen too late
•
Communication gaps lead to missed interventions
Social Workers
•
Spend 3+ hours daily on admin instead of direct care
•
Manage scattered notes across multiple systems
•
Experience emotional burnout from constant high-stakes decisions
Young People
•
Need someone to talk to "in the moment" – not next week's appointment
•
Want judgment-free conversations without formal processes
•
Often need support outside traditional office hours
User Interviews
Field Persona Sketch – Mary McAbbit, 15, Walthamstow
User Interview – NHS Doctor, Nivedita Boss
Stakeholder Interviews
Stakeholder Interview – Jeremy Pitt (Recording Notes)
Stakeholder Interview Template – Jeremy Pitt, Founder
Discussion Guide – Social Worker Interviews
Competitor Matrix
Assumptions Mapping
Target Audience Mapping
Research Approach
I conducted in-depth interviews with social workers, young people, and NHS professionals, using AI tools to synthesize insights at scale. Used Notably AI to cluster qualitative data from 50+ interviews, GWI Spark to validate findings with global behavioral data, and journey mapping to identify critical pain points.
Understanding the Problem
The breakthrough came from a simple observation during my research: Young people need help at 2 am. Social workers can't be available 24/7. But what if AI could echo their care and presence during those crucial gaps?
Echo was born from this insight: AI shouldn't replace human empathy; it should amplify it.
"I wish I had more time to actually be there for the kids."
— Social Worker Interview
The Spark: Why Echo?
Echo works on two levels, providing complementary support for both young people and social workers.
For Young People
Crisis Support
Intelligent routing to human help when needed
Non-judgmental Space
AI trained to listen without triggering formal processes
24/7 Availability
Safe, empathetic chat companion when crisis hits
For Social Workers
Streamlined Workflow
All information in one place, 40% faster case prep
Risk Intelligence
Color-coded alerts highlight concerning patterns
Smart Summaries
AI-generated case updates save 2+ hours daily
For the System
Data-Driven Insights
Understanding patterns across cases
Resource Optimization
More time for human connection, less for admin
Early Intervention
Risks flagged before they become crises
The Solution: Echo AI Assistant
Design Process
1. Research & Synthesis
Mixed-methods research combining human interviews with AI-powered analysis. Used GPT-4 to analyze interview transcripts, identifying patterns manually missed.
2. Ideation & Concept Development
AI-assisted brainstorming using Miro AI to structure ideas, then human validation with SMEs. Echo as "empathy amplifier" rather than replacement.
Crazy 8s ideation sketches & storyboard concepts
Persona (Aisha, Social Worker) · Problem Statement · Storyboard · User Testing
3. Prototyping with AI
Used GPT-4 and Claude to simulate Echo's responses, testing tone and safety protocols without building full technical infrastructure.
We built a working prototype to validate the concept with real users.
Try the live product built during the research phase →
Try the Prototype
4. Design & Testing
Parallel design streams with constant cross-validation between user groups. Mock conversations with AI agents to simulate real usage patterns.
Key Features & Design Decisions
Dashboard for Social Workers
Design Philosophy: Essential information at a glance, with drill-down for details.
TL;DR Section: AI-generated case summaries
Risk Snapshot: Color-coded alerts (green/amber/red)
Next Steps: Intelligent task prioritization
Smart Notifications: 50% reduction in alert noise
Chat Interface for Young People
Design Philosophy: Familiar messaging experience with enhanced safety.
Empathetic Tone: Copy refined through AI testing for warmth and safety
Visual Cues: Subtle indicators when conversations are being monitored
Crisis Protocols: Seamless handoff to human support when needed
From early wireframes to high-fidelity prototypes — the screens below show Echo's evolution across both the social worker dashboard and the young person's interface.
High-Fidelity: Social Worker View
Echo's AI-powered case management dashboard — risk snapshots, AI summaries, follow-up plans, and session transcripts.
AI Generated Follow-up Plan
AI Summary Audio & Transcript
AI Generate – TL;DR Summary
Child Profile – About Tab
Case Dashboard – Risk View
Case Dashboard – Overview
Wireframes
Social Worker Flow — Case dashboard, messaging, profile & reporting screens
Young Person Flow — Home, Echo chat, scenario responses & follow-up
Mobile UI Screens
Qualitative Impact
"Finally feels like someone's listening, even at 3 am."
— Young Person
"I can focus on being present instead of note-taking."
— Social Worker
Earlier interventions, fewer crisis escalations
— System Impact
Quantifiable Outcomes
3+ hours
Saved daily per social worker (validated through user testing)
60%
Faster identification of high-risk situations
40%
Reduction in case preparation time
Proof of Concept
Built interactive prototypes that demonstrated AI conversation flows, tested them with 20+ social workers, and validated the time-saving projections through workflow analysis.
Results & Impact
Key Learnings
AI Amplifies, Doesn't Replace
The most successful AI solutions enhance human capabilities rather than substituting them. Echo works because it makes social workers more effective, not redundant.
AI-Powered Design Process
Using AI throughout research, design, and testing accelerated development while maintaining human-centered outcomes. AI tools became collaborative partners, not just production aids.
Safety Through Intelligence
In sensitive contexts, AI can actually improve safety by consistently monitoring for risks that exhausted humans might miss.
No child unheard. No worker unsupported.
Echo represents a new paradigm in social care technology, where AI doesn't just automate tasks but genuinely supports human connection. The project showed that when we design AI to complement rather than compete with human empathy, we can create tools that make care systems more responsive, workers less burned out, and young people truly heard.
Watch the Demo Video
bottom of page