This project develops an AI-enabled digital onboarding platform to enhance clinic efficiency, consultation flow, and clinical documentation in a hand orthopaedic clinic. A defining feature of the initiative is that it is designed and led by a clinician for clinicians, ensuring that the system aligns closely with real-world clinical workflows, time pressures, and documentation requirements—addressing gaps often observed in technology solutions developed without direct clinical ownership.

The platform integrates AI automation using n8n, with an AI agent powered by the DeepSeek reasoning API, delivered via Telegram as the patient-facing interface. It functions as a pre-consultation onboarding tool, allowing patients to describe their symptoms and clinical history at registration. Importantly, the platform serves as an impartial digital interface, enabling patients to articulate their concerns in their own words without time pressure or perceived judgement. This often results in more complete and candid symptom descriptions than those obtained during time-constrained consultations.

Patient inputs are then summarised into a structured clinical format, which clinicians can review prior to consultation. This supports shorter and more focused consultations, reduces repetitive questioning, and facilitates more efficient note-taking and medical record documentation after the consultation. By focusing on onboarding rather than real-time consultation capture, the platform complements existing clinical interactions and avoids the limitations associated with AI tools that attempt to transcribe or interpret doctor–patient dialogue during consultations.

The system is deployed using a clinician-in-the-loop, human-guided approach, with clinicians retaining full oversight and control over how information is reviewed and applied. AI functions as a workflow assistant rather than a decision-maker, supporting safe, trusted, and practical integration into everyday clinical practice.

The project follows a two-phase evaluation approach, beginning with simulated testing to assess handling of multilingual and informal patient inputs, followed by deployment with real patients in the clinic. Evaluation focuses on consultation time savings, usability, and enhanced doctor–patient interaction, by allowing clinicians to focus on the essence of the patient’s concerns rather than administrative or repetitive information gathering.

Capacity building is a core outcome of the project, supported by a team of young researchers from clinical backgrounds trained in AI evaluation, testing, and refinement. Looking ahead, the platform is designed for integration into routine clinical workflows and scaling across other orthopaedic subspecialties and clinical domains. The team is actively seeking private-sector partners to support system maintenance and scale-up, enabling sustainable and scalable deployment.