Skip to content

Work

The problem, the calls I made, and what the numbers did.

Every figure below comes from a system that ran in production for a real client of Everglade Systems. Three builds are written up in full, the way I'd walk a hiring manager through them. The rest are in one table.

In depth

HealthcarePrivate AIComplianceDiscovery

A specialist physician group, 30+ physicians

$450K

in compliance fines avoided, March to August 2026

A private, air-gapped AI for a 30-physician group: $450K in fines avoided in five months

A specialist physician group was missing compliance deadlines and paying for it, and couldn't use any AI that touched the internet. I led discovery, scoped a private on-site AI to three jobs, and defined the metrics. Five months after launch, $450K in penalties simply hadn't happened.

My role: Discovery and problem definition, scope and prioritization, success metrics

Timeline: Live March 2026; results measured through August 2026

Read the full write-up ›
Law FirmAI AssistantDocument OCRPrioritization

An immigration and personal injury firm, 6 attorneys and 12 paralegals

+$30K

more revenue every month within six months, same 18 people

An in-house AI assistant for an 18-person law firm: $100K to $130K a month without hiring

An immigration and personal injury firm was stuck at $100K a month because growth meant hiring, and every question in the office went to the same overloaded paralegals. I found the real bottleneck, scoped an assistant that had read every file, and measured the result: 50 hours a week back and 30% more revenue in six months.

My role: Discovery and problem definition, scope and prioritization, success metrics

Timeline: Six months from launch to the $130K month

Read the full write-up ›
Real EstateAI PipelineGrowthMetrics

A real estate company

$2.3M

in new revenue over six months, about 75% of it profit

One AI pipeline for every lead: $2.3M in new revenue for a real estate company in six months

Leads arrived from four sources and each source lost a few. I mapped where leads died, scoped a single pipeline that handled every lead the same way, and set the metric that mattered: leads that became closed deals. Up 35% in six months, worth $2.3M at roughly 75% profit.

My role: Discovery and problem definition, scope and prioritization, success metrics

Timeline: Six months from launch to the $2.3M figure

Read the full write-up ›

In brief

Seven more builds. Same rule: real numbers from live systems, and the part I owned.

Large healthcare clinic Problem: Hand-typed patient and insurance details caused rejected claims and late fees. Shipped: ML check on every claim before submission, inside their EHR. 400+ billing hours/week back; $340K+ in fees avoided My role: Discovery, scope, success metrics
B2B sales team Problem: Call prep took hours, so it only happened sometimes. Shipped: Research agent that briefs the rep 15 min before each call. 80% less prep time My role: Problem definition, spec
Coaching practice Problem: Every client submission waited on one person to grade it. Shipped: Computer-vision app scoring against the coach’s own checklist. 99% auto-scored; waitlist gone My role: Scope, acceptance criteria
Film production company Problem: Loyalty tracking in three forms and a hand-updated spreadsheet. Shipped: Automated tracking, milestones, and follow-ups on their existing tools. 300%+ customer growth, no data entry My role: Discovery, prioritization
Service business Problem: $1,500/month for a few SEO articles nobody read. Shipped: AI content system the team edits instead of writes. 95% lower cost; 4x output; 8–16x traffic My role: Scope, metrics
B2B company Problem: Daily LinkedIn presence kept falling off the list. Shipped: Autonomous posting, research, and competitor-monitoring agent. 1M+ impressions; 95% lower cost My role: Problem definition, metrics
B2B sales team Problem: Outbound depended on one busy person remembering to follow up. Shipped: Automated connection, follow-up, and meeting-request sequences. 500+ decision-maker connections/month My role: Scope, sequencing decisions

About the missing names

Clients aren't named. Some work in fields where privacy is the whole point, and some folded these systems into their own offerings. Every number was measured against how things ran before. Some of the work was delivered with trusted partners under Everglade's direction. Happy to say more in conversation.

Want the parts that didn't make the page?

Every one of these has a decision I'd defend and one I'd take back. I'm glad to go through them.