Context
FastFinder helps molecular diagnostics labs analyse PCR runs. The users are lab specialists who review hundreds of results a day under accreditation rules; the business needed the product to scale from a single assay tool to a suite that enterprise labs would buy.
Business context
- Goal
- Win enterprise lab contracts.
- Metric
- Demo-to-contract conversion.
- Trade-off
- Showing the AI’s power in a demo versus giving lab specialists the caution and transparency they need to trust it. Resolved by making confidence and overrides visible in the same screen the demo runs on.
What made it hard
- Trust: an AI call is only useful if a specialist can see why it was made and overrule it in seconds.
- Volume: hundreds of samples a day, multiple plate configurations. The review screen had to work at keyboard speed.
- Scope creep for good reasons: inventory, consumables and traceability were operational needs the lab could not separate from analysis.
- Two kinds of lab, one product: highly automated labs with instruments driving the workflow, and manual labs where every step is done by hand. The same screens had to serve both.
- Regulation: every change had to be explainable in an audit and consistent across the whole suite.
What I did
Research in the lab
Observed and interviewed specialists across multiple labs; mapped the review workflow from plate load to signed report.
Results-review interface
A dense, keyboard-first table with confidence, flagging and bulk approval patterns that keep the human in charge.
End-to-end inventory tracking
Stock, consumables and lot traceability designed into the same system, so a result can always be traced to its reagents.
Design system and demo flow
A component library for the product suite, and a redesigned demo path that let sales show value in twenty minutes.
The result
For the business: the product opened the door to new markets, with the redesigned demo path carrying it into enterprise deals. For the lab: full digital traceability of every sample, instrument and consumable, from order to signed report.
- A multi-step run wizard keeps a workflow that spans hours and days on one rail: the specialist always knows where they are and what comes next.
- Input samples and the 96-well output plate stay visible on the right through every step, so assignments can be checked without leaving the task.
- Volumes are checked against what the step needs and updated automatically in inventory, so traceability is complete without extra data entry.
Reflection
I would start with the highly automated labs: fewer steps, a simpler workflow and less human intervention make for a faster go-to-market and a cleaner proof of concept. Support for smaller, manual labs can follow once the core is proven.