Hi, I'm Tony —

I research, prototype, ship & build AI products from 0 → 1.

AI generalist, senior UX researcher and product designer with 10+ years shaping AI-powered products in ambiguous, early-stage environments — across UK private healthcare, fintech and crypto. I turn foundational field research into things that actually get built.

◦ London, UK ◦ AI generalist ◦ Senior UX researcher ◦ 0→1 product strategy

Selected work

Products I've researched, shaped and shipped.

Most of these are live demos — click through and have a look around.

CodeCatalyst ↗ · an AI operations layer for UK private healthcare (2024–present)

Ava

An AI receptionist that answers every call, reads the live calendar, books appointments in real time, and catches complaints mid-call to escalate them safely.

Voice AI · booking · escalation

Pulse

A patient lead-generation and pipeline tool — every enquiry tracked to a booked appointment, with capacity across doctors and automated follow-ups.

Pipeline · CRM · automation

ClaimBridge

The billing-and-claims side of the ops layer — connecting practice workflows to clean, trackable claims so nothing slips between treatment and payment.

Claims · billing

The Bridge

A CQC-aligned complaint system that turns issues detected during a call into a structured, compliant handling flow.

Compliance · CQC

ConceptNet · founder & CEO — voice IP infrastructure for AI

ConceptNet

A voice-native agent OS that turns enterprise voice commands into intellectual property companies can own, stack and license — positioned as a cheaper, faster, 200-language alternative for voice classification and intent. Live sandbox on Hugging Face, open-source classifier on GitHub, and benchmarks against GPT-4o.

Voice-native OS · 200 languages · open-source · currently raising

How I work

Research first, then build to think.

01 Get in the room

For CodeCatalyst I spent weeks embedded in a real private practice — recording calls, interviewing the practice manager and staff — to map how bookings, complaints and compliance actually work before designing anything.

02 Research that ships

I use field insight to define product strategy and MVP scope, not just slides. The point of the research is to change what gets built — and to be honest about what to cut.

03 Build to think

I'm a daily generative-AI user — LLMs and voice agents for prototypes, workflows and synthesis. I iterate between prototype and live demo fast, and build Chrome extensions in my spare time to stay sharp.

04 Ambiguity is the job

Early-stage, 0→1 environments are where I'm most useful. I'm comfortable defining a problem space from scratch and moving before everything is certain.

Experience