Everett Considine — Give me the problem

Two careers, one vessel

The seams are gold.

I spent the first half of my career as a programmer, architect, and consultant — Bellcore, IBM, Pfizer, Amway's Quixtar, a startup CTO seat, my own ecommerce company. I built systems for a living and rode every technology wave by building on it.

Then I spent fifteen years on a different kind of system: people. As a Level 3 Internal Family Systems practitioner, I've guided more than 7,000 people through behavior change — and learned that what looks broken in us is usually something trying to protect us.

AI joined the two halves. The join isn't a scar — it's the strongest, most visible line in the work.

Read the full story →

A deep teal vessel mended with gold kintsugi seams — stronger and more beautiful at the joins.

1998 – 2011

The builder

Enterprise ecommerce architecture, consulting from Fortune 50 to five-person shops, company owner, startup CTO. A complete online store in five weeks. Forward-deployed before the title existed.

2011 – 2025

The human systems years

IFS parts work, The Peaceful Eating Way, co-author of Rezoom, co-builder of Self-Therapy Journey — psychology as interactive software, a decade early.

2025 – now

The reunion

Directing AI as my implementation layer: a production app with attack-verified security, governed agentic automation, attribution systems, an instrumented revenue engine.

How I work

Verification isn't a phase. It's the method.

Diagnose the real problem

What's actually happening — underneath what people say is happening.

II.

Prototype rapidly

Working software in days: AI as the implementation layer, thirty years of judgment as the editor.

III.

Test against reality

Live probes, real transactions, synthetic users. A success response proves the call succeeded — not that it did what you meant.

IV.

Secure and govern

Attack my own fixes. Keep humans in the loop where stakes are real. Design for the failure, not the demo.

Ship, document, iterate

Complete releases, living documentation, honest accounting of what's still imperfect.

The thesis

AI can tell people what to do. It can't see why they don't.

Humans aren't one voice — we're many. One part wants change; another fears it; another rebels; another just wants relief. Most AI treats that inner conflict as noncompliance. A parts-informed behavioral intelligence layer would recognize it — and respond in ways that widen a person's choices instead of narrowing them.

I've spent fifteen years mapping how people actually get stuck, and I've already shipped the first primitive. The rest is buildable.

"If the AI becomes dramatically more effective — is that also good news for the human using it?"

Give me the problem. I'll build the system.

Built, tested, governed, shipped.

People trust it with their most vulnerable moments.

Every client gets their session — automatically.

$25,000 out of the gate.