What’s changing in women’s health - and what still isn’t

Women’s health is getting more attention than it has in decades. New products, platforms, and funding categories are emerging under the femtech umbrella. But beneath that progress, long-standing gaps remain - in research, in clinical practice, and in how health systems are designed.

This was the starting point for a conversation hosted by Shen AI, bringing together three perspectives from across the space: Romain Schneeweis, co-founder of AINOHA; Amy Keenan, Chief Marketing Officer at Hello Inside; and Dr. Vanessa Hamoniaux, a hormone specialist based in Brussels.

‍

The scale of the gap

The imbalance in women’s health is well documented, but it continues to shape how products are built today. Less than 2% of total VC investment goes to femtech (1), products serving half the population. Only around 7% of healthcare research focuses on conditions exclusively affecting women. (2)

The implication is structural. AI models trained on existing health data inherit that bias. Algorithms built on datasets that skew toward male populations will produce outputs that do not fully reflect women’s physiology - across hormones, metabolism, and immune response.

As Amy Keenan pointed out, this is a relatively recent shift:

‍

“Women have only been included in clinical trials for about 30 or 31 years. And sex as a biological variable in scientific studies has only been included since 2016.”

‍

The result is not just a research gap, (3)(4) but a system-level bias that continues to influence how women’s health is interpreted and addressed.

‍

Data that actually helps

One of the central tensions in women’s health technology is how data is used. Many products focus on tracking - symptoms, cycles, biometrics - but without interpretation, more data does not necessarily lead to better understanding. In some cases, it increases uncertainty.

Romain Schneeweis described this clearly:

‍

“It’s not about delivering more data. It’s about understanding patterns over time - across weeks and months - in a way that gives useful context for both the user and their doctor.”

‍

The emphasis is on patterns over time, not isolated data points. At Hello Inside, this translates into a focus on relevance and usability. As Amy Keenan explained: “It’s not just about sharing data. It’s about getting information back that helps you understand your lifestyle and what’s actually influencing your health.”

The shift here is from data collection to decision support, helping users understand what is happening in their bodies and why, in a way that is actionable.

‍

Personalization at scale with constraints

AI has made it easier to personalize health experiences, but the discussion highlighted an important nuance: personalization is only useful when it is grounded in reliable data and clear boundaries.

One of the risks is that models can reinforce existing assumptions rather than challenge them. As Amy Keenan noted:

‍

“When you're talking to a model, it’s reinforcing your belief system, and that can apply to health as well.”

‍

At the same time, these systems are often trained on datasets that do not fully represent women: “At the moment, that’s a biased dataset; typically a white, able-bodied man in the US". This creates a dual challenge: ensuring personalization is both accurate and appropriate.

The approach described in the discussion is to introduce constraints, combining AI capabilities with clinical oversight and longitudinal data. This allows personalization to scale, while maintaining a level of control that is necessary in health contexts.

‍

Designed around the wrong baseline

A deeper issue underlying many of these challenges is how healthcare systems were originally designed. Women’s health has often been treated as a variation of male health, rather than a distinct domain with its own complexity.

In practice, this leads to gaps in diagnosis, interpretation, and care.

Dr. Vanessa Hamoniaux sees this regularly in her clinical work:

‍

“Women often feel misunderstood. Their symptoms are frequently minimized or explained away - not because they aren’t real, but because the research hasn’t fully caught up with how those symptoms present.”

‍

This becomes particularly visible in cases where symptoms are complex or span multiple systems. Patients may go years without a clear diagnosis, not because the symptoms are unclear, but because they are not being connected.

As she describes:

‍

“Sometimes it takes 10 or 15 years before everything comes together. When we review the full set of symptoms, many women realize for the first time that they are not separate issues, but part of the same underlying problem.”

‍

Technology has the potential to improve this by helping connect disparate signals, but only if it is designed with these differences in mind.

The trust problem

Building trust in women’s health is not just a UX challenge. It reflects a broader history of being dismissed or not fully heard.

Dr. Vanessa Hamoniaux sees this in her clinical work: “We need to take the full set of symptoms seriously and revisit them consistently, not dismiss them. Too often, women feel they are not being heard”.

For technology to earn that trust, it has to operate differently. Romain Schneeweis emphasized the importance of restraint:

‍

“We made the deliberate choice early on not to promise answers that we cannot deliver. We don’t say we will tell you what’s wrong. We say we’ll help you understand what’s happening.”

‍

Amy Keenan added a systemic perspective on trust in AI-driven systems:

‍

“Trust shouldn’t be assumed by default. It’s important to understand what a healthcare app is doing and how your data is being used.”

‍

The implication is clear: trust is not built through features alone, but through transparency, validation, and clearly defined boundaries.

‍

The bigger picture

What stands out in this discussion is not just the scale of the problem, but the clarity with which it is now being described.

There are still gaps: in research coverage, in medical training, and in how healthcare systems are structured. But there is also increasing alignment around what needs to change.

Romain Schneeweis summarized this shift:

‍

“There are so many issues in women’s health, and just as many ways to help. But more importantly, it’s not about looking, it’s about listening, because this is happening right in front of us.”

‍

The problems are visible. The demand is real. And the technology to address them is becoming more capable. The remaining question is not whether the gap exists, but how intentionally it is addressed.

‍

Sources:

(1) UNICEF Office of Innovation, Investing in Femtech

(2) World Economic Forum, Prescription for Change: Policy Recommendations for Women’s Health Research

(3) NIH Revitalization Act of 1993 

(4) Fatima Farrukh MD, Richard C. Becker MD, American Heart Journal (volume 287, Sept. 2025), Sex as a biological variable: a contemporary perspective

‍