Artificial intelligence, real results.

I founded Lina Saúde in 2019 to help health insurers and medical teams look after chronic patients. In practice, nurses followed each patient closely. In the beginning, everything happened over the phone.

The problem

After every session, the nurse had to reread the whole conversation and manually log everything into the medical record, in a separate system. Repetitive, slow, error-prone work that stole time from actual care.

What changed

I redesigned the operation end to end. I moved care from phone calls to WhatsApp, where patients already are. And I put an AI, adopted in 2022 in the early days of the ChatGPT API, in charge of the invisible work: after each session it sweeps the full conversation (text, audio, photos and documents), extracts what matters and logs everything into the patient's medical record. The nurse types nothing.

What this teaches

The 10x gain did not come from AI alone. It came from understanding the process and redesigning it with technology and AI together. It's the thesis that guides me at MCA Advisory to this day.

• Nurses about 10x more productive after the operation was redesigned, from phone calls to WhatsApp with AI
• Half of the nurses' time given back to care through automated medical records
• 1.5 million active lives followed by the platform
• Acquired by Conexa in November 2024, a digital health ecosystem backed by Goldman Sachs, General Atlantic and Pátria

Read the full case

Conexa is one of the country's largest digital health ecosystems, backed by Goldman Sachs, General Atlantic and Pátria. After the Lina acquisition, I moved through product areas and became director of artificial intelligence. Three products came out of that tenure.

AI Scribe: the whole appointment without typing a word

Physicians used to spend part of every appointment typing notes instead of looking at the patient. AI Scribe changes that: the appointment happens naturally and, at the end, the AI processes the full audio and structures everything into the medical record. I built it in-house and put it head to head with several market solutions. Ours won every time.

AI support: much lower cost, higher satisfaction

The path was method, not magic: first I dug into patients' most common problems (passwords, connecting with the doctor, prescription questions), and only then came automation, running in the app and on WhatsApp, with handoff to a human whenever needed. The result breaks the myth that automating support degrades experience: costs fell substantially and satisfaction went up.

Conexa Aí: healthcare entirely inside WhatsApp

A single channel where patients talk about their health, reach a doctor right away, book appointments, ask about their prescribed medication and get preventive care guidance based on official Brazilian Ministry of Health practices. Before routing to a doctor, the AI understands the case. The doctor walks into the appointment already knowing what the patient needs, and both sides win.

• AI Scribe adopted by every physician in the operation, used every day
• Support AI that resolves most tickets without a human
• Substantially lower support costs, with NPS and satisfaction going up

Read the full case

Latta was born from a pain I've lived with for four years as the owner of Nacho, a shih tzu with a sensitive belly: finding the right product, remembering vaccines and meds, and that nagging doubt about whether a vet visit is needed. Pet owners fend for themselves between pet shops, clinics and Google.

What Latta does

A single channel on WhatsApp where the owner talks freely and the AI handles the rest: purchases at the best price with reliable delivery, well-reviewed services near home, diet or symptom questions answered on the spot. No app to install.

Why this case matters for your company

Neither I nor my partner Lucas Bergman are trained developers. I've always worked in tech with a CTO and a dev team by my side. At Latta, I'm the one building everything: journey, infrastructure, security, scalability and database, with AI-assisted development. It's proof that, with the right method, business people can build and operate a real product. And that's exactly what I teach.

• Complete product (journey, infrastructure, security, database) built with his partner Lucas Bergman, without a development team
• Public launch on August 6, 2026

Try Latta now on WhatsApp

Read the full case