The use of data, patient-centred care and meaningful stakeholder engagement by healthcare providers has already unlocked significant benefits for underserved patients, and helped many organisations to build a better, clearer picture of the populations they serve – enabling them to target interventions more effectively. Yet, with the rise of AI, bias remains a key challenge to overcome, often built into services either by design or by a lack of relevant training data. How can we anticipate and overcome drivers of inequality, and what frameworks and processes are relevant as we continue to design more equitable healthcare systems - for practitioners and patients.