AI for good could finally make women’s health visible at work
For 25 years I worked in women’s health, much of it in the NHS. The women I saw rarely came in to talk about work, but work was always in the room.
I met nurses who’d cut their hours because they couldn’t trust their memory on a long shift. Women who turned down promotions in the year their symptoms became unmanageable. Women diagnosed with ADHD in their forties, after decades of masking it at work. None of them appeared in any workforce data as a women’s health story. If they showed up at all, it was as sickness absence, reduced hours or a resignation.
In 2024, a report by the NHS Confederation, the CREATE Health Foundation and London Economics estimated that women missing work because of painful periods, endometriosis, fibroids and ovarian cysts costs the UK economy around £11 billion a year. It also estimated that about 60,000 women can’t work because of menopause symptoms.
Those are big numbers, and almost certainly an undercount. They cover a handful of conditions. They don’t capture what happens when hormonal change collides with neurodivergence, chronic pain or burnout. And they can’t see the women who never take a sick day but quietly cap their own careers.
Health data and work data live in different places. A GP record knows about symptoms. An HR system knows about absence. Neither knows how the two connect for one woman, across a month, a year or a career.
Often women don’t connect them either. If you’ve been told for years that it’s stress, you stop expecting an explanation. When my team surveyed 620 women about their wellbeing, one clear theme was how hard it was to put into words what was happening to them, let alone explain it to a manager.
This is where I think AI has a real role, and it’s less about clever diagnostics than about patterns.
A woman’s experience is spread across ordinary days: sleep, focus, energy, mood, pain, sensory load, how work felt. AI is good at spotting patterns in that kind of everyday information, without judgement. Used well, it could help a woman see her own picture and walk into a GP appointment, or a conversation with her manager, with something concrete to say.
Shared with consent and anonymised, the same information could show employers and policymakers something they’ve never had: how women’s health shapes working lives, beyond a single absence code.
None of this works unless a few things hold. The data has to belong to her first, with consent that’s real and value that comes back to her before anyone else.
Support has to stay clear of diagnosis, because helping someone understand herself is different from telling her what’s wrong. Employers have to use what they learn to adapt work, not to screen people out, or women will rightly stop sharing. And it has to work on the hard days, for women with the least capacity, not just the most motivated.
If you lead people, ask what your data can’t see. If you build AI, involve the women it’s meant to serve from day one. If you work in policy, treat women’s health as a productivity issue, not a niche benefit.
We already know women’s health costs the economy billions. What we’ve lacked is the evidence to act on it, woman by woman and workplace by workplace. For me, that’s what AI for good looks like: not replacing care, but making visible what’s been missed for far too long, so women can be heard, understood and supported to keep working.
Carly Walter 83 in the Foundever AI 100 UK 2026
