The Use of Artificial Intelligence to Support Triage and Navigation in the NHS

What is the issue?

Timely access to the right care remains a central NHS priority, with triage at first contact playing a key role in directing patients according to need. This process now operates across a diverse and increasingly complex landscape, combining human-led assessment, telephone and in-person routes, and digitally-assisted tools. The growing use of digital and AI-enabled triage reflects ambitions to improve efficiency, manage demand, and enhance patient experience, alongside policy interest in integrated “digital front door” models. However, this expansion introduces complexity in how different triage approaches interact, how decisions are made, and how patients navigate multiple access pathways.

Realising the potential benefits of AI-enabled triage is challenging. Current evidence highlights variability in tool performance, as well as the importance of trust, user behaviour, and clinical oversight in shaping outcomes. Implementation is further constrained by issues such as data quality, interoperability, and the need for clear governance and accountability frameworks. Concerns about equity also persist, including unequal access to digital tools and biases in underlying data. Despite rapid development, the evidence base—particularly in primary care and first-contact settings—remains limited, making it difficult to assess impact and guide effective adoption.

How are we helping?

The DECIDE Centre will carry out the following:

• Map the current and emerging landscape of AI-enabled triage in primary care, including technologies, suppliers, use cases, and policy drivers.
• Analyse how AI is currently being used in triage and navigation at first contact, including when and why it is adopted or not, and its effects on workflows, access pathways, and patient experience.
• Examine stakeholder perspectives (patients, clinicians, policymakers, suppliers) on what constitutes effective and equitable AI-enabled triage, and what “good” could look like in practice.
• Identify opportunities, risks, and challenges associated with AI-enabled triage, including safety, equity, infrastructure, and implementation barriers—particularly in relation to integration with the NHS App and wider system architecture.
• Develop future-facing visions and practical pathways for designing, embedding, scaling, and sustaining AI-enabled triage within the NHS.
• Co-produce evidence-based recommendations, tools, and resources to support policymakers, practitioners, and patients in navigating and shaping AI-enabled triage systems.
• Centre patient and public perspectives, including lived experiences and issues of access, inclusion, and digital capability, through targeted engagement and co-design activities.
• Synthesize findings across multiple workstreams into actionable outputs to inform policy, practice, and future research.

The evaluation adopts a rapid, developmental and formative approach, combining multiple methods to capture both current practice and future possibilities. This includes literature and landscape reviews, qualitative interviews (including in-situ “go-along” methods), and stakeholder workshops using visioning and back-casting techniques, alongside co-design with patients and communities. Analytical frameworks such as NASSS, digital candidacy, and sociotechnical systems perspectives are used to understand complexity, implementation, and equity. The key aim is to generate a holistic, system-level understanding of AI-enabled triage—capturing not only how it works, but for whom, under what conditions, and how it can be developed in ways that are effective, equitable, and sustainable.

Project protocol

The project protocol is available xxx