AI Quality Engineering & Communication
Quality without clarity
is just noise.
A career spent in testing — passionate about quality in every form. But here's the uncomfortable truth: a perfect test run means nothing if the person who owns the outcome can't understand what it's telling them.
Now fast-forward to AI. Quality is harder to measure. Communicating it is harder still. Business owners are being asked to trust systems they can't see, built on processes they can't follow, producing results they can't interpret.
ThatsQuality.ai exists to close that gap — translating AI quality into the language of confidence, prediction, and business outcomes. Not test reports. Not dashboards full of numbers. A clear verdict. An answer they can act on.
Bringing it to life
What I'm building right now
Three connected products. Each tackles a different slice of the same gap: making AI quality legible to the people who own the outcome.
AI projects
AIQIDE
The narrative engine. Turns raw eval results into a per-persona story of business impact, autonomy, rollback, and regulatory weight.
Crystal Ball
The dashboard. Pulls signals from Langfuse and other eval tools into one verdict view per persona, with drill-down into the trace.
Scout
The autonomous AI quality exploratory testing agent. Generates hypotheses, runs probes, surfaces findings against a target system.
Note: demo email addresses must be whitelisted. Please try your work domain when prompted.
AI tools
Ganglia
A multi-agent thinking system. Specialised agents research, structure, and attack a hard question, while a coach sharpens your own reasoning in parallel.
Galah
Push-to-talk dictation that runs entirely on your device. Speak anywhere; Galah transcribes locally and tidies the text with AI. Your audio never leaves the machine. Windows, Mac, Android, iOS.
Currawong
A meeting recorder that transcribes, identifies who spoke, and writes the summary — all on-device. Recordings and voiceprints never leave your phone.
Poppy
An AI chief of staff that does the work, not just the chat: drafts email, schedules, runs research, and ships projects, with a memory that persists across every session.
Meridian
Persistent memory for AI agents. A structured notes store agents read and write so context survives the session, the machine, and the model — the layer the rest of the suite is built on.