An HL7 bridge and report builder for a diagnostics chain

A cloud platform for a diagnostics operation, with an HL7 bridge that brings instrument messages in from the laboratory floor and serverless functions that process them. Report layouts are arranged by drag and drop rather than hard-coded, generative AI assists with drafting, and device fingerprinting is used to tie sessions to known machines.

The problem

Instruments spoke HL7 on the laboratory network while the platform the business wanted to run on was in the cloud. The two could not reach each other, and report layouts were changing often enough that every change becoming a development task was itself the bottleneck.

What we built

1. 01

   A bridge process on the laboratory network that receives HL7 messages locally and forwards them to the cloud, so instruments keep talking to something on their own segment.

2. 02

   Cloud functions to process the forwarded messages, so handling scales with volume instead of running on a machine in the laboratory.

3. 03

   A drag-and-drop report builder, so a layout change is an afternoon's configuration rather than a release.

4. 04

   Generative AI used for drafting inside the workflow, where it removes typing rather than adding a chat box.

5. 05

   Device fingerprinting to recognise the machines sessions come from.

What it does now

• Laboratory instruments feed a cloud platform without being exposed to the internet.
• Report layouts are changed by configuration, not by a deployment.
• Message processing scales with volume rather than with one machine's capacity.