Home / Use cases
Use cases

Eight kinds of hospital AI, filed apart

An AI scribe and a sepsis model carry very different risks. We file every deployment by what it does, and note what buyers and patients should check.

2 sourced entries

Radiology & imaging

Software that flags urgent findings on scans, prioritises worklists or supports reporting for X-ray, CT and MRI.

What to checkCE marking class under the MDR, sensitivity for the findings it flags, and how alerts reach the radiologist.
0 sourced entries

Pathology

Analysis of digitised tissue slides to detect, grade or measure disease.

What to checkValidation on local slides and scanners, and how results are signed off by a pathologist.
No sourced entries yet. Know of one? Submit it.
0 sourced entries

Early warning

Prediction of sepsis or patient deterioration from vital signs, labs and notes.

What to checkAlert rates, false alarms, and whether outcomes were measured after go-live.
No sourced entries yet. Know of one? Submit it.
0 sourced entries

Triage & patient flow

Prioritising emergency and outpatient demand, and predicting no-shows or admissions.

What to checkBias across patient groups, and who can override the tool's priority.
No sourced entries yet. Know of one? Submit it.
0 sourced entries

Decision support

Suggestions for diagnosis or treatment shown to clinicians at the point of care.

What to checkIntended purpose, human oversight design, and EU AI Act high-risk obligations.
No sourced entries yet. Know of one? Submit it.
0 sourced entries

Patient communication

Drafting letters, answering patient messages and supporting follow-up.

What to checkReadability, escalation to a human, and data protection for patient messages.
No sourced entries yet. Know of one? Submit it.
0 sourced entries

Operations

Scheduling, bed management, capacity planning and coding.

What to checkMeasured effect on waiting times or capacity, not only staff time.
No sourced entries yet. Know of one? Submit it.