0-to-1 AI Product · Built with Lovable

SpeakEasy Order — Voice-Guided Food Ordering

A voice-first food-ordering assistant for elderly and low-vision users — replacing small text, dense menus and multi-step taps with a guided, spoken interaction that confirms each choice out loud before committing.

Role
Solo builder — problem to working prototype
Method
Problem discovery, voice-flow design, prompt engineering
Scope
Accessibility, voice-guided ordering
The problem

Food-ordering apps were designed for people who don't need help ordering

Small text, dense menus and multi-step taps are a minor annoyance for most users and a real barrier for elderly or low-vision ones. The fix isn't a bigger font — it's an interaction that doesn't require reading a screen at all.

Solution

SpeakEasy Order: spoken, confirmed, and never rushed

GUIDED VOICE

A conversation, not a menu

The assistant asks what you want to eat the way a person would, instead of expecting you to scan a grid of small photos and prices.

CONFIRM ALOUD

Every choice read back before it's placed

Each item is confirmed out loud before moving on, so a misheard word gets caught before the order is placed, not after it arrives.

NO RUSH

Paced for the person, not the interface

There's no hidden timer or auto-advancing screen — the flow waits for a spoken answer instead of assuming a fast tap.

Voice greeting Hear order Confirm aloud Place order
Prototype

A working app, not just a flow diagram

SpeakEasy Order is built and live — the guided voice flow and spoken confirmation work end to end in the prototype below.

Live prototype — SpeakEasy Order Open full screen ↗
Open interactive prototype ↗
Process

Started from a real accessibility gap, not a feature list

Discovery
Problem discovery — used Claude to pressure-test who actually struggles with app-based ordering and how often, before designing anything.
Flow
Voice-flow design — mapped the minimum spoken interaction — greeting, order, confirmation, placement — with no screen-reading required.
Prompts
Prompt engineering — wrote and iterated the conversation design that carries the voice interaction naturally.
Build
Prototype — shipped a working, clickable app on Lovable rather than leaving the concept as a deck.