Skip to content

IPSO

In development

Voice AI assistant for self-service laundromats: customers scan a QR on the machine, explain the problem out loud and the agent fixes it or escalates.

My role

Product owner and lead developer: scope, architecture, data model and a 60+ task delivery plan; implementation delegated to Codex and reviewed by me against tests.

Problem

When a machine fails there is nobody on site; customers lose money and owners find out late.

Solution

A no-install web app opens from a QR, holds a voice conversation, runs admin actions on the machine (unlock door, retry cycle) through the IPSO Connect integration, logs every incident and emails the owner, who reviews details and transcripts in a dashboard.

Highlights

  • Multi-tenant data model with row-level security
  • Machine simulator for safe end-to-end testing
  • Built with a planned, test-gated Claude Code + Codex workflow

Architecture

IPSO — Architecture (in development) A no-install web app opens from a QR, holds a voice conversation, runs admin actions on the machine (unlock door, retry cycle) through the IPSO Connect integration, logs every incident and emails the owner, who reviews details and transcripts in a dashboard.
Back to projects