From paper and guesswork to a live view of every asset.
Chenot's lab was spending up to two days interpreting each gene analysis by hand, holding up the doctors who needed the results. Cloudbliss built a Power Apps tool that ingests the machine exports, sends them to an AI with a proprietary prompt and knowledge base, and returns a structured health-profile report. The bottleneck is gone.
Up to two days to interpret each gene analysis by hand — a bottleneck on the doctors waiting for results.
Chenot's lab interpreted each gene analysis manually, a process that took up to two days per case. That capped the lab's throughput at the speed of a person working through each interpretation, and it held up the doctors who needed the results to advise their patients.
A Power Apps tool that ingests the machine exports and hands them to a governed AI with a proprietary prompt and knowledge base.
Cloudbliss built a Canvas app that ingests the gene-analysis machine exports and sends the data — alongside a proprietary prompt and a knowledge base — to an AI for interpretation. The AI analyses how genes interact to produce a structured report on the patient's health profile and suggested pathways to improvement; the lab runs the tool, and qualified doctors use the report to advise patients.
It runs on a SharePoint back end with an HTTP integration to the AI (the client manages their own AI API, keeping running costs in their control) and a RAG implementation for continuous learning. Dynamic inputs are passed through with the prompt — additional patient context, medications that may affect results, customisable gene templates and configurable thresholds — so the platform adapts through configuration, not code.
It was delivered by two developers and a technician, with the COO overseeing, in roughly 40 hours of focused build.
Analysis time cut from two days to ten minutes — without taking the clinician out of the loop.
Interpretation dropped from up to two days to about ten minutes. At around 20 analyses a week, that is roughly 300+ hours a week returned versus the manual process, with faster turnaround for doctors and, in turn, their patients.
The platform is flexible and configurable — new templates, thresholds and context are handled through configuration rather than a rebuild. AI output is treated as probabilistic and supports rather than replaces clinical judgement: reports are used by qualified doctors, with the human-in-the-loop sitting firmly with the clinician.