The biology of in vitro fertilization (IVF) will never be simple. Fertility treatment depends on countless variables, every patient responds differently, and no technology can remove that uncertainty. What technology can change is everything around it.
After working with thousands of fertility patients during my career, I noticed that many of their biggest frustrations had little to do with medicine itself. The system often adds its own burden: travelling two hours for a fifteen-minute appointment, taking time off work for routine monitoring, organising daily life around a clinic’s schedule. People took unpaid days off or even left their jobs to be able to do fertility treatment.
That observation became the foundation of Plan Your Baby. Instead of asking patients to keep returning to one central clinic, we built a telehealth-first model supported by more than 450 partner locations across the UK. Most of the monitoring and consultation can now happen closer to home, while specialist care remains centralised. Today, around 85% of the patient pathway can take place close to where patients live. That means less travel, less time away from work and fewer practical disruptions during treatment.
The same principle now shapes how we think about AI. The AI hype has made it easier than ever to build and fund software around healthcare. However, healthcare already has plenty of dashboards, apps or isolated tools. What it often lacks is something that connects clinicians, diagnostics and decision-making in a way that actually works in practice. In that environment, existing care providers have an important advantage over AI-first start-ups: they already understand how things work, where time is lost and what patients need. They can use technology to scale what they already do well.
Clinicians spend a lot of time gathering information, reviewing previous treatment, checking results before they can apply clinical judgement. If technology can reduce some of that work safely, it can give doctors more time for the decisions that genuinely require their expertise.
At Plan Your Baby, we are developing an AI model to help prepare cases for clinicians rather than make decisions for them. We want the system to analyse each patient’s complete clinical history, laboratory results, highlight the information that matters most, and then organise it into a structured case for the physician. The treatment strategy remains the doctor’s responsibility.
Our system is being trained on anonymised data from more than 20,000 patient journeys collected through years of fertility care. Records are de-identified before being used for model development. The aim is to combine medical knowledge with patterns drawn from real clinical experience while protecting patient privacy.
In practice, the technology should help clinicians see the full context more quickly: what has already been tried, which results matter, what may need closer attention and which questions still need answering. Over time, it may also help identify patterns associated with how treatment is likely to progress.
Interestingly, greater personalisation can also lead to greater standardisation. Medicine will always involve clinical judgement. Ask twenty experienced fertility specialists how they would approach a difficult case — and you may hear twenty slightly different answers. What technology can do is make sure each clinician starts with the same complete information and a consistent view of the evidence before applying their own judgement.
Any system used in this way also has to meet a much higher bar than an ordinary AI product. We are developing it with the appropriate medical device certification in mind from the outset. In healthcare, the technology must satisfy the same standards expected of any other clinical tool.
The conversation around AI in healthcare is evolving. We have spent the past few years exploring what foundation models are capable of. In the next phase, we should focus less on what the models themselves can do, and more on building products that solve real clinical problems. That requires not only technical excellence, but also years of domain expertise, deep understanding of clinical workflows and the ability to turn AI into something healthcare professionals can actually use.
