Neighborhood Guide: West Town, Chicago
A walkable neighborhood guide with a live map, walking times and real-time transit.
Open the live appWHAT WE'VE BUILT
What it looks like
when operations stop
requiring you.
These are first-party systems — designed and deployed inside our own operations before we built for anyone else. The architecture is public. The methodology transfers. External client work is not published without engagement.
Live demos
Real apps we built and run. Open one and use it.
CASE 01 — MOVEMENT SOLUTIONS
What it looks like when a practice
stops running on its founder.
A boutique physical therapy practice running on founder presence. Every clinical decision, every communication exception, every scheduling conflict routed to the owner — not by choice, but by default.
There was no system. There was just Lance.
WHAT WAS BUILT
A multi-agent AI operating system: twelve named agents, a 470-rule clinical routing engine, a client communication system, a content pipeline, and a governance layer — built on Google Cloud Platform (GCP) with a HIPAA Business Associate Agreement, Vertex AI, Gemini Gems, and Claude Projects.
THE STACK
Google Cloud Platform — GCP (Vertex AI — HIPAA BAA)
Gemini Gems + Claude Projects (multi-agent routing)
Next.js + Supabase
Constitutional governance layer (470-rule clinical router)
Deployed: Vercel · movement-solutions.com
WHAT RUNS NOW
Clinical routing, session preparation, client communication, content strategy, and consulting intake now operate with the founder as approver — not executor.
Nothing auto-sends.
Everything passes through a human gate before it reaches a client.
THE ARCHITECTURE
ARTIFACT — System Architecture Diagram — Multi-Agent Ecosystem
Agent-to-agent relationship map, autonomy tiers, and data flow through the governance layer — without exposing clinical data or client records.
CASE 02 — THE MOSO APP
A clinical intelligence layer
built from first principles.
Pattern recognition was happening in the clinician's head across sixty active clients.
Session-to-session continuity depended on memory —
a resource that doesn't scale and doesn't transfer.
The gap wasn't skill. It was the infrastructure.
WHAT WAS BUILT
A web application routing interaction logs through a 25-module clinical communication framework — structured around human behavioral patterns, adaptive communication styles, and nervous system state recognition — to surface longitudinal client patterns and generate session preparation briefs before each appointment.
THE STACK
Google Cloud Platform — GCP (Vertex AI — HIPAA BAA)
Next.js + Supabase Auth
Gemini 2.5 Flash / Pro + Claude (best-fit model routing)
Task-based matrix router (Claude ↔ Gemini switching)
Agent cost monitoring
Constitutional governance layer (clinical inference boundary)
Deployed: Vercel
WHAT RUNS NOW
The clinician enters a session with a brief already generated — cross-referenced against the client's 6-month history, nervous system state patterns, and prior session progression.
Memory is no longer the bottleneck.
APPLICATION SCREENS
Today — session brief view
Client monitoring dashboard
Screens are shown in a private walkthrough. No client records or patient information appear.
THE FRAMEWORK
ARTIFACT — 25-Module Clinical Communication Framework Map
Visual architecture of the clinical intelligence layer — module relationships, routing logic, and output structure.
CASE 03, FIRST OUTSIDE BUILD
In progress.
The first outside build to complete the Plan and Build steps. The write-up posts when the client approves it. Include a request in your intake form if you want to be notified when this posts.
If you want to understand what we build before applying, these cases show the methodology at full resolution. Ask for the architecture diagrams in your intake form and I will walk you through them.
Work with us →QUICK ANSWERS