Hi Reader,

Between 2020 and 2024, Amanda Ducach kept her company in stealth for four straight years.

Over that stretch, companies kept coming to her, checkbooks open, hoping to advertise through her product. Easy revenue, sitting right there for the taking. She turned it down every time.

That decision is significant because Amanda is CEO and co-founder of Ema, the AI for women's health, and Ema exists to do one thing: Drive a positive health outcome for the woman on the other end.

The moment a company pays to have Ema recommend their product over a better-fit option, Ema stops serving the user and starts serving the advertiser. So she declined, for four years, and built a proprietary hybrid language model instead of cashing in on the consumer momentum she already had.

That discipline paid off in a way investors can count. Out of stealth, she closed over a million dollars in enterprise contracts inside roughly six months, in a category the market barely knew existed. Her work has been covered in the Wall Street Journal and Forbes, and Ema won FemTech World's 2025 AI Innovation of the Year.

I've taught more than two million people how to work with data and build with AI, and I spend my days as a fractional CMO helping technical founders turn real capability into a sale. So the part of our conversation that stuck with me was Amanda's most challengingproblem today. It sits far from whether the product works. It's that she's so far ahead of the market she has to teach buyers the category before she can sell inside it.

​Read the full interview >>​

WHAT WE DISCUSSED

Why she kept the company in stealth for four years and walked away from easy ad revenue to protect the product's core job.

How following the data from the very first MVP let her team catch signals that clinicians and other users were missing.

Why over eighty percent of interactions inside Ema stay in her own infrastructure, because open models still fail on women's health questions more than sixty percent of the time.

The astronaut and the suit: how she makes a complex hybrid model legible to a buyer in one picture.

Why her hardest problem today is teaching the category, and the pre-call move that turns a vague demo into a specific outcome the buyer can see.

How narrowing to a handful of ideal customer profiles, with the right demos built for exactly them, became her growth motion.

WHERE TO FIND AMANDA DUCACH

​Ema​

​Amanda Ducach on LinkedIn​

MY BIGGEST TAKEAWAYS

• The clearest signal of a strong founder in this conversation was the four-year no. Most founders under runway pressure take the ad money and tell themselves they'll fix the incentives later. Amanda understood that the incentive is the product, and she guarded it. That single decision shaped everything that came after.
• Her data rigor is the through-line. Ten years ago, on the first version of the product, her team was watching the data closely enough to catch early signals of a crisis that the textbook version would have missed. Plenty of founders say people will want this. Far fewer can point to the data that proves it, and she always could.
• The astronaut metaphor is worth stealing. A frontier model is powerful and helpless in space until you add the suit. Her product is the suit: the guardrails, the safety layer, the contextualization that lets an AI work in a health setting. She kept the full depth of the product and made it legible in one image, which is exactly the skill most technical founders are missing.
• Her hardest problem is the tell of someone operating ahead of the curve. When your biggest obstacle is that buyers are still learning to picture the category, you've built something truly new. Her answer is to shrink the vast capability down to one specific use case the buyer can see themselves inside, and the sale closes itself once they can.
• The move I kept coming back to is engineering the moment before the demo. Most buyers struggle to name their own use case on their own. So you build a lightweight step ahead of the conversation that pulls it into the open, then you hand them the exact picture. That's the difference between a demo that lands half the time and one that lands closer to always.

If you're a technical founder and your pipeline feels thinner than your product deserves, this one is worth your time.

All the best,

Lillian Pierson

Fractional CMO & GTM Engineer

​
​