Paula Bellostas Muguerza, global lead of healthcare and life sciences at Kearney, and Elena Bonfiglioli, global business leader for healthcare, pharma and life sciences at Microsoft, warn against using skewed data to improve women’s health.
Women’s health has long been underfunded, under-researched and designed around the male body, with structural consequences that leave women spending around 25% more of their lives in ill health. Growing calls to redesign healthcare with women in mind have made one thing increasingly clear: closing this gap will depend not just on policy and funding, but on how data and technology are collected, built and used.
As technology and AI become part of everyday healthcare, there’s a real opportunity to address long-standing gaps, using better data to improve research, diagnosis and care. But without the right action now, there’s also a risk those same tools could reinforce existing bias and exclusion for years to come.
The foundations are weak. Women make up half of the global population, yet health systems have rarely been designed around their lived experience. Only 7% of global healthcare research funding goes to conditions that exclusively affect women, and just 5% of medicine is properly tested and labelled for use during pregnancy or breastfeeding. As a result, women’s health remains systematically underrepresented across research, clinical trials and real-world data.
When AI is trained on this skewed base, bias scales quickly. Triage tools can underweight symptoms such as pain or fatigue, while diagnostic algorithms may perform less accurately for women, particularly in cardiovascular disease, stroke and cancer risk assessment.
Generative AI raises the stakes. With around 24% of UK adults already seeking health advice from AI tools or social platforms, emerging evidence shows these systems are more likely to downplay symptoms in women and ethnic minorities. At scale, this risks quietly shaping care prioritisation and access in ways that are hard to detect and even harder to undo.
This is not a failure of technology, but the predictable outcome of building powerful systems on incomplete foundations.
Rebuilding the foundations
Technology alone cannot undo decades of exclusion, but it can help prevent those failures from being repeated. AI systems trained on robust, sex-disaggregated data can surface patterns that were previously invisible. This includes how symptoms present differently, how treatment responses vary, and how risk accumulates across the life course. Standardised data frameworks and interoperable platforms allow information to move with patients rather than remain trapped in organisational silos.
The principle sits at the heart of the [w]Health Tech Manifesto, which calls for women’s health data to be treated as core infrastructure, not a specialist add-on. Tools like this translate this ambition into practice, enabling organisations to assess where gaps persist across the health value chain and prioritise action.
Data only matters if it leads to better care. Yet, women’s health pathways remain deeply fragmented. From menstruation to maternity and through to menopause, women’s care pathways are inconsistent, poorly connected and chronically under-resourced. The consequences are already visible: gynaecology is now the largest specialty on NHS waiting lists, accounting for 12% of patients aged 18-64.
Digital platforms offer a way to redesign end-to-end care. Virtual consultations, remote monitoring and AI-supported triage can help prioritise patients more effectively, identify risk earlier and reduce unnecessary referrals. When embedded into redesigned pathways, rather than layered onto existing ones, these tools can ease pressure on overstretched services while improving access and outcomes.
The same opportunity exists in maternity care, where digitally enabled pathways can replace episodic touchpoints with continuous, coordinated support across pregnancy, birth and postnatal care. Women’s health hubs show what is possible when services are designed around women rather than specialties; scaled with shared data and digital tools, they could finally deliver joined-up care from menstruation through menopause.
AI as an enabler, not a shortcut
AI can be a powerful enabler in women’s health. Used well, it can support clinicians by analysing complex datasets, identifying risk earlier and tailoring interventions more precisely. It can strengthen research through more inclusive trial design and support women directly with personalised information and navigation that reduces delay and uncertainty.
AI cannot operate in isolation, and clinicians must remain firmly in the loop. Human oversight is essential to validate outputs, protect patient safety and ensure accountability. Without it, automation bias risks displacing clinical judgment, eroding trust and widening disparities rather than closing them.
Addressing these risks goes beyond technology alone. It demands clear guardrails on how data is collected to truly serve 51% of the population and creates better outcomes for women. Sex-disaggregated data must become standard, legacy datasets must be scrutinised, and there needs to be collaboration across private and public sectors.
Technology will not fix women’s health by default. It will only do so if it is designed, governed and deployed with intent and with women involved from the outset. The tools already exist. The risk is not slow progress, but fast adoption on the wrong foundations. The choices made now will determine whether AI finally closes the women’s health gap or locks it in for another generation.
