Jason Stiles.

I help enterprises turn ambitious customer experience ideas into systems that actually work.

I've spent 25 years helping large organizations modernize how they engage their customers: shaping strategy, architecture, and delivery, and guiding initiatives from concept to production at scale.

Today I lead a cross-functional team responsible for an enterprise customer-experience platform that uses machine learning and predictive intent to handle 150k+ customer voice interactions a day across dozens of locales. I work hands-on with applied ML and agentic AI, and I build my own products to explore where the technology is heading next. My favorite work sits at the intersection of customer experience strategy, emerging technology, and practical delivery: figuring out not just what's possible, but what should actually be built.

Projects

DialogWright
An open-source TypeScript framework for voice and chat agents, easy enough to run yourself and powerful enough for a contact center. A decision model understands what callers mean, code decides what happens, and a policy gate checks every action. It grew out of Jev IVR.
Hedge
Calibrated trivia: a daily AI-assisted mobile game built to explore modern AI development workflows end-to-end, from prompt engineering and agent tooling to large-language-model integration. Live on the App Store.

Building Hedge: Lessons from Shipping AI Software →
Jev IVR
A phone line you can talk to, with no LLM underneath. A mixed-initiative voice IVR built on TypeSafe's Jev model: callers speak naturally, the code stays deterministic, and every decision is a typed question with a probability you can test like software. Demo video and open source.

Writing

Essays
Machines of Loving Grace, Two Years On: reading Dario Amodei's 2024 essay about what powerful AI could do for the world, graded from the floor of the industry now that some of its bets are being scored.