In this article:

Bookable uses our internal AI matching algorithm to take the request a patient makes and match it to a clinician capability or other category. It takes into account all of the information the patient provides, including their date of birth and sex at birth.

We then recommend which pathway a patient should follow to receive care. These pathways may be external NHS services (Dental, Self-referral for pregnancy care) or internal appointments with particular clinicians (Nurse, ANP, GP).

We determine which pathway to recommend to the patient based on the clinician capability or other category that has been matched.

We provide this recommendation as an option, and allow the patient to ignore the recommendation to book with a GP if they feel that is more suitable.

The overall flow looks like:

✅ Request → Categorisation → Recommended pathways → Patient choice

Patients submit a free‑text request in Bookable, plus date of birth and sex at birth. We use this to assign the request to a clinician capability or other category so it can be routed appropriately.

A clinician capability may describe a symptom, condition, or procedure. They are aligned to defined competencies and typical patterns of demand in primary care.

Clinician capabilities are not exhaustive. They represent presentations that allied health professionals (AHPs) can safely and appropriately manage in clinic.

Requests that fall outside these clinician capabilities are intentionally routed to a GP. This reduces inappropriate AHP bookings and supports safe care navigation.

This mean here are some very specific clinician capabilities, like ‘Smear test’, and some very broad ones like ‘Skin - Rash’.

This is because without much more information (or even a picture) it’s not possible to know what kind of rash the patient is presenting with, they just need to see someone who can make an assessment, diagnosis and mange their rash, or escalate appropriately.

A smear test, on the other hand, is very easy to be specific about. For example, the patient might write: “I need to book my smear test”.

We prioritise safe allocation over exhaustive matching. Ambiguous presentations default to recommending the patient book a GP appointment.

Clinician capabilities and routing rules are reviewed and may change over time to reflect practice needs, workforce capability, and local pathways.

We are confident in our matching based on the information patients provide. However, if you believe a request was miscategorised or inappropriately allocated to an AHP appointment, please flag it (by email for now to [email protected], please give us as much information as you can).

Feedback informs category tuning and safety rules. We refine our matching so requests reach the right clinician first time, while keeping the patient experience simple and seamless.

We have audited the internal matching algorithm across a large internal test set of real patient requests. We ensure that the accuracy is high for these test cases, and that the fallback mechanisms are safe in the case of a mismatch.

The care navigation only provides a recommendation of a next step or a clinician to book an appointment with, and does not block the patient from booking with a GP if they prefer. Because of the limited scope it is not a medical device at this point.

We have carried out a full clinical safety review, as part of our DCB0129 for Bookable with care navigation. The documentation for this, including the hazard log, is available on request.

We have written a self-serve guide to run you through the steps.

You will need to add some slots for your non-GP roles, wait 20 minutes for us to scan your appointment book, and then configure their roles in the Hub via "⚙️ Settings" > Your practice > "Bookable". See more about setting up non-GP roles in the Hub.

If you want to learn more about why we built this, and the difference between Care Navigation and Clinical Triage, see our blog written by the Product team!