AI in healthcare covers very different kinds of work. Drafting an inspection report, researching cancer biology, understanding practice finances and coordinating appointments each require different evidence and oversight. A useful starting question is simple: what task is changing, and how will we know it has improved?
The Best Application of AI in Healthcare category at AI Awards 2026 brings together Mental Health Commission, Nexomic, Sláinte Finance and SPRYT.
The profiles below draw on the official AI Awards 2026 finalist summaries. Organisation links provide further background. The applications span administrative, research and operational work; they should not be treated as interchangeable clinical tools.
Meet the finalists
Mental Health Commission
The Mental Health Commission uses supervised generative AI to draft regulatory reports from inspection findings. Structured inputs, tailored prompts and expert validation support the reporting process while retaining professional accountability.
Visit the Mental Health Commission.
Nexomic
Nexomic’s DeepOmic combines multiple types of biological data to investigate cancer subgroups, treatment response and metastatic risk. Built on AtlasX, its explainable biomarkers are undergoing prospective validation.
Sláinte Finance
Sláinte Finance connects GP practice income, banking, staffing and clinical information to identify unclaimed entitlements. AI helps categorise and explain information, while calculations remain deterministic and auditable.
SPRYT
SPRYT’s Asa supports appointment booking, reminders and patient queries through familiar messaging channels. Conversational AI and no-show prediction aim to improve access and reduce administrative work.
Different applications need different measures
AI Ireland’s editorial recommendation is to assess each workflow against its intended purpose. The following questions are for leaders considering similar projects. They are not an evaluation of the finalists’ performance.
Reporting: measure the finished document
For an AI-assisted report, assess the final approved version. Compare the original findings with the draft, record omissions and corrections, and include review time in any efficiency measure. A faster first draft only helps if the overall process preserves accuracy and gives reviewers a clear basis for approval.
Research: make the validation stage clear
Ask what the research is intended to establish and which questions remain open. Separate exploratory findings from results tested on new data or through prospective studies. When communicating a project, describe its current stage accurately. Recognition for research should not be presented as proof of clinical benefit.
Practice operations: check the numbers and their source
For a financial workflow, ask whether someone can trace an output to the underlying records and reproduce the calculation. Test incomplete entries, duplicates and corrections. Agree who reviews a proposed action before it affects practice decisions. Measure the checking work as well as the convenience of asking a question.
Appointments: test the whole journey
For appointment support, look beyond the number of messages sent. Review whether people can book, cancel or reschedule successfully and reach a member of staff when needed. Include people with different access needs in testing, and retain an alternative route for those who cannot use the digital channel.
Start with a clearly bounded pilot
A leadership team can make the next step concrete by agreeing five things before a pilot begins:
- The task: Exactly what the tool will assist with, and what sits outside its scope.
- The owner: Who reviews outputs and handles exceptions.
- The information: Which records are needed and who is authorised to access them.
- The measures: The starting position, quality checks and outcomes that matter.
- The review: When the team will decide whether to continue, change or stop.
The aim is a clear account of what changed and whether that change helped. Keep operational measures, research findings and clinical outcomes distinct so that the evidence supports the claim being made.
Recognition and evidence
AI Ireland organises the AI Awards. This article combines published finalist summaries with editorial analysis. Shortlisting is not independent verification of performance, clinical effectiveness or regulatory approval.
Explore the full AI Awards 2026 shortlist and our article on responsible AI and ethics.
See practical AI in action
Join us at AI Awards 2026 to celebrate the organisations and people putting AI to work.
Want to bring these lessons to your organisation?
Discover more from AI Ireland
Subscribe to get the latest posts sent to your email.
