Telehealth, diagnostics, and pharmacy. Run as one system.
A protocol becomes adherence. Adherence becomes P&L. I build the layer that makes that happen.
Fractional or full-time COO. I build the backend operating layer for US telehealth and diagnostics companies scaling from $5M to $150M. Run with AI: a production agent on a live customer channel recovered about 45% of leads lost to response lag. Less lag, pharma-grade compliance, optimized EBITDA.
Let's compare notesThree moves, in order, every time (that changes everything)
See the board
I connect the dots most people miss (the seam between operations, AI, clinical, and revenue that specialists don't see because each is looking at one square).
Invent the system
I design the operating model the problem actually needs, not the one that's convenient.
Build it
I don't advise on operations. I build the machine that runs them, then I make it survive contact with real volume.
One career across every square of this board. That's how I see it whole.
Operations built from zero, at scale
Built the patient-operations engine for an AI telehealth company from nothing (SOPs, KPI dashboards, intake and follow-up workflows, cross-functional accountability).
Read the case study →AI, implemented with discipline
A production AI agent on a live customer channel (Claude API + ERP + WhatsApp) with scoped escalation, confirmation gates, auto-pause on human reply, and an append-only audit trail.
Read the case study →Synthesis, turned into commercial result
Delivered regional growth at Omron LATAM on a 7,000-point distribution network, and negotiated multi-market exclusive licensing.
Read the case study →Proof by What I've Built
An AI Sales Agent on a Live Channel: The Guardrails Are the Product
A DTC medical-device operation was losing close to half its warm leads to reply lag. I put a production AI sales agent on the live WhatsApp channel — Claude API into Odoo — to qualify, quote from the live ERP, and write orders. But the agent was never the hard part. The guardrails are the product: a price-integrity rail, human-in-the-loop takeover, and one append-only audit trail, all behind a single middleware. Thirteen days to production, running 24/7.
Omron LATAM: Product Development & Go-to-Market Machine
Omron LATAM had product but no go-to-market machine built for the region's specific gaps. I designed the launch framework and distribution model from scratch, then built a 7,000-point POS network and DTC channels to run it through. Fifty-plus product launches later, the result was 300% regional growth and the #1 market position, with multi-market exclusive licensing negotiated on top.
HelixVM: Patient-Ops Engine from Zero
An AI telehealth operation was treating conversion as a number to hit. I rebuilt it as a variable to steer: a stage-by-stage conversion cascade that located every drop and its cause, which revealed the real constraint was capacity, not intake. Built the instrument before the fix.
IRIS / FENIX: The regulated operating system, built from zero
A founder-led pharma and diagnostics business with no operating infrastructure. I built the ERP, the quality system, the SOPs, and the pharmacovigilance function as one system, and carried it through GWDP certification.
Insights
You can't afford a COO? The math just flipped
Why AI turned senior operating leadership from a late-stage luxury into a survival hire, for small companies most of all.
Read the article →AI Isn't a Great Leveler. It's a Great Un-Masker.
The "AI helps low performers most" story is real, and it may be measuring the wrong thing.
Read the article →We only have two settings for new technology. Both are wrong.
What the dot-com crash teaches us about where AI is, and who wins from here.
Read the article →Every metric is a tributary. Only one is the river.
The AI-native operator connects every function, but everything must point at one thing. In a world where capital is no longer easy, that thing is cash.
Read the article →The F1 Business Era
Once every team rents the same AI, the edge shifts from big product leaps to compounding micro-adjustments, lap after lap.
Read the article →How to Give an AI Agent Memory Without Breaking the Audit Trail
Why agent memory should be a projection of the same audit log, not a separate notebook, one model summaries can never promote to fact.
Read the article →About
I'm Hugo Acurio, a healthcare operations leader based in Fort Lauderdale (bilingual English/Spanish).
I've spent fifteen years building the operating layer that turns technology into patient and provider outcomes: the patient-operations engine behind 100K+ encounters at an AI telehealth company, a production AI agent now running on a live customer channel, and 300% regional growth at Omron across LATAM. I also founded and operate a regulated pharma and diagnostics business (full P&L, distributor networks, regulatory affairs, a cross-functional team).
The thread through all of it: I see the seam others miss, I design the system, and I build it to last.
Let's Compare Notes
If you're a founder or CEO scaling a DTC health, telehealth, or diagnostics company and the operating layer still needs building, reach out.
See where your operation leaks