What Industrial AI Taught Me About Health & Wellness
My earlier AI work focused on data-rich operational businesses. Today, I am applying those lessons to health-and-wellness brands and creators through General Dev.
The target market changed. The discipline did not: start with a valuable business problem, build the smallest useful system, and measure what changes.
Start With the Economics, Not the Model
“We should use AI” is not a product strategy. A useful build starts with one economic outcome:
- Revenue growth — help more customers find the right product, program, or next action.
- Margin expansion — remove repetitive work from a specific, measured workflow.
- Product ideation and creation — turn trusted expertise or proprietary data into something customers can use and pay for.
That focus keeps the work grounded. The model, interface, and integrations follow from the outcome rather than becoming the point of the project.
What Transfers to Health and Wellness
Operational AI taught me that reliability matters more than novelty. Health-adjacent systems raise the standard further because trust can be lost quickly.
- Approved sources: The system should work from reviewed content, product data, and business rules — not improvise health claims.
- Clear consent: People should understand what data is being used and why.
- Human review: Experts remain responsible for consequential decisions and exceptions.
- Visible measurement: Conversion, retention, time saved, and quality should be defined before the build begins.
- Clear boundaries: The product should say what it can do, what it cannot do, and when a person needs to step in.
These are not compliance decorations added at the end. They shape the product from the first prototype.
Three Wedges Worth Testing
1. A Guided Customer Journey
Help a customer find the right meal plan, membership, course, or coaching program using an approved catalog and explicit business rules. This is more useful than adding a generic chatbot because it is tied to a decision and a measurable outcome.
2. A Delivery Copilot
Give a coach, practitioner, or creator a system that organizes approved knowledge, drafts routine follow-up, and prepares the next action for human review. The goal is to reduce a defined workflow’s manual effort without removing judgment.
3. Expertise as a Product
Package proprietary methods, content, or data into a focused customer-facing experience. A strong AI product does not imitate the founder. It makes the founder’s best work easier to access and apply.
Proof Before Philosophy
The useful lessons came from shipping real products:
- Fuel Meals: an adaptive meal-matching experience for a Shopify Plus business.
- Randall Reilly: a conversational product built on proprietary industry data.
- Idias Health: a healthcare-financing product I am building with a partner in Mexico City to help patients cover eligible out-of-pocket care, with approved cases paid directly to providers.
- Linear BJJ: a voice-first AI training journal designed to help athletes reflect and train with more intention.
Fuel Meals and Randall Reilly are commercial proof. Idias Health and Linear BJJ are founder-built health products. Their roles are different by design: Idias focuses on healthcare financing; Linear BJJ uses AI to support athlete reflection and training.
How I Start
The first engagement is an AI Leverage Blueprint: three days to identify the highest-value wedge, model the economics, pressure-test the available data, and leave with a prototype plus a fixed pilot scope.
The standard is ambitious but explicit: target 10× value relative to project cost, validate the assumptions, and instrument the system from day one. Targets depend on the baseline, data, and operating constraints; they are not guarantees.