AI voice recognition · Speech therapy · Child-centred UX
Making speech practice easier to track without interrupting the therapy session.
Wombat Words was an AI-assisted speech therapy support tool created to help therapists track word practice during sessions with children. Instead of asking the therapist to manually count every word attempt, or adding another person into the room, the product used voice recognition to listen for target words and show progress in a simple, glanceable way.
Wombat Words was an AI-assisted speech therapy support tool created to help therapists track word practice during sessions with children. Instead of asking the therapist to manually count every word attempt, or adding another person into the room, the product used voice recognition to listen for target words and show progress in a simple, glanceable way.
Speech therapy with children rarely happens in a neat, controlled environment. Practice can happen while reading, playing, moving around, or responding naturally in conversation. Therapists needed a way to understand whether target words were being practised enough without breaking the flow of the session or shifting their attention away from the child.
I worked as the UI/UX designer, using Figma to explore the interface, structure the experience, and prototype the flow. My research and planning included direct conversations with the therapist HOD to understand how sessions worked in practice, what needed to be tracked, and how much attention the interface could realistically ask from the therapist. Prototype testing sessions were run with the therapist HOD, myself, and the project manager.
The interface needed to be useful without becoming another task. It had to support quick progress checks, target-word visibility, and session awareness, while staying simple enough for a therapist to use during a live, child-led therapy environment.
I focused the experience around calm, glanceable feedback — helping therapists see what still needed practice without pulling them away from the child. The product needed to support the therapist quietly in the background, not become the centre of the session.
Wombat Words showed how AI could support a very human workflow. Voice recognition helped reduce manual counting, but the real product value was in making therapy sessions feel less interrupted, more natural, and easier for therapists to manage in the moment.