AI Diabetes Clinic: AI-First Clinic Model for Type 2 Diabetes Diagnosis, Therapy Optimisation, and Complication Surveillance, with a Saudi Arabia Implementation Blueprint

Diabetes is one of the fastest-growing global health emergencies. The International Diabetes Federation (IDF) estimates that in 2024 there were 589 million adults aged 20–79 years living with diabetes worldwide (1 in 9); 43% (252 million) were undiagnosed; diabetes caused 3.4 million deaths; and global health expenditure reached USD 1.015 trillion, with 81% of adults with diabetes living in low- and middle-income countries [1 ]. The downstream burden is driven by preventable complications: diabetic kidney disease affects roughly 30–40% of people with diabetes; diabetic foot ulcers occur in an estimated 19–34% over a lifetime; diabetic peripheral neuropathy affects about 30% of people with diabetes globally; and diabetic retinopathy affects about 22% of adults with diabetes (with ~6% having vision-threatening disease) [2–5]. Saudi Arabia is a high-burden setting within the Middle East and North Africa region: IDF 2024 estimates indicate 5.34 million adults (20–79 years) living with diabetes, an age-standardised prevalence of 23.1% (approximately one in five adults), and 43.6% undiagnosed. This combination of high clinical need and rapidly maturing digital health infrastructure makes Saudi Arabia an ideal implementation testbed for an AI-first, end-to-end care model that can be generalised internationally [1]. This Personal View proposes a fully AI-enabled, clinician-in-the-loop “AI Diabetes Clinic” that synthesises patient-entered data, laboratory and imaging results, and longitudinal records to support diagnosis of type 2 diabetes, personalised pharmacotherapy selection and titration, and continuous complication surveillance. The model combines an AI-assisted pre-visit synthesis, an explainable treatment optimiser aligned with contemporary guidelines, and automated screening workflows for retinopathy, nephropathy, neuropathy, and diabetic foot disease, with escalation to clinicians when risk thresholds are crossed. Because AI-first clinics concentrate clinical, privacy, and cybersecurity risks, deployment must be coupled to robust governance: model transparency, bias and safety testing, audit trails, and human oversight; privacy-by-design and consent aligned with local regulations; and cyber-resilience aligned with national standards. We outline an evaluation strategy spanning usability, clinical effectiveness, equity, safety, and cost-effectiveness, including continuous post-deployment monitoring to ensure sustained performance across populations and care settings.