AI-enabled ultrasound service
Point-of-care ultrasound that reaches the patient record.
Health systems are buying handheld ultrasound faster than they can account for it. Elephant runs the ultrasound service around the probe, from the scan at the bedside to a documented, reviewed and coded record. The software layer underneath is ePOCUS, built by our sister company, Dragonfly Technology.
The problem
The scan is the easy part. Everything after it is where programmes fail.
Probes arrive through donor programmes, national procurements and hospital budgets. What arrives with them is rarely a way to see what they are doing. Scans sit on the device or in a siloed app, findings never reach the patient's record, and nobody can say which probes are in use, which health workers need support, or what the programme actually changed.
Ghost scans
Examinations performed and never documented. No record for the next clinician, no evidence for the funder, no basis for a claim.
No line of sight on quality
Without structured review, nobody knows whether an image was adequate or an interpretation was safe until something goes wrong.
Fragmented systems
Every deployment meets a different mix of record systems, insurers and billing rules. Custom integration work stalls rollouts for months.
Connectivity assumed, not present
Software that needs a connection to function is software that stops at the district boundary. Capture has to work offline and reconcile later.
How it works
From capture to claim, in a single pass
Each stage happens once, at the point of care, and produces structured data the next stage can use.
Capture
The clinician scans on a handheld probe. Images and study metadata are taken at the moment of capture, whatever the device.
Document
Indication, views obtained, findings and next steps are recorded at the bedside in the flow of the exam, not retyped later.
Quality check
Studies are routed to a reviewer, scored against agreed criteria and attributed to a named clinician, building a credentialing trail as the service runs.
Coded record
The study is written to the patient record as structured, standards-based data rather than a flat image or a PDF.
Claim
Where a payer exists, the documented study carries the codes it needs, so the work performed is the work reimbursed.
Stages three and five are the two that programmes most often leave out. Dragonfly documents both in detail: POCUS quality assurance and POCUS billing and revenue capture.
The software layer
Any probe, any record system, any payer
ePOCUS is middleware. It sits between the ultrasound device and everything the health system already runs, so a deployment does not wait on custom integration work at every site.
Device agnostic
Works with handheld and cart-based probes from different manufacturers, including hardware a programme has already bought.
Interoperable by design
Built on HL7 FHIR with standard clinical terminologies, so studies land in ElephantOS or a third-party EHR as structured records.
Offline first
Scans are captured and documented without a connection and sync when one returns. Connectivity stops being a condition of deployment.
Open to AI tools
Third-party ultrasound AI can be attached to the workflow, with its involvement in each study logged rather than assumed.
Auditable end to end
Every scan, review and amendment is attributed and time-stamped, giving programme owners an evidence trail rather than a headline number.
Real-time programme view
Activity, quality scores and referral outcomes by probe, clinician, facility and district, available as the programme runs.
Who we build this for
Two different questions, one service
Health facilities and hospital groups
- Documentation happens during the exam, so clinicians finish a scan and move on.
- Findings reach the patient record where the next clinician will look for them.
- Supervisors can review a junior colleague's studies without standing beside them.
- Where a payer exists, completed studies carry the coding needed to be reimbursed.
Governments, ministries and funders
- A live count of scans performed, by facility and by district, not an end-of-programme estimate.
- Visibility of which probes are in active use and which are sitting idle.
- Quality scores and credentialing progress by health worker, so training goes where it is needed.
- Reporting that feeds existing health information systems instead of standing beside them.
Track record
Built on public health systems already running on our software
ElephantOS instruments the facility. ePOCUS instruments the diagnostics. ePOCUS is developed by our sister company Dragonfly Technology, a clinician-led engineering team that came out of the same frontline deployments. More on the people behind it on the Dragonfly Technology team page.
Common questions
What people ask before a deployment
What is ePOCUS?
Do we have to buy probes from Elephant?
Will it work with our existing electronic health record?
What happens where there is no reliable connectivity?
How is scan quality assured?
Can ultrasound AI be used alongside it?
Planning an ultrasound programme?
Tell us what hardware you have, what record system you run, and what you need to be able to prove. We will show you how the service would work in your setting.
