Last week’s Linkedin post about the $800k mass spec running 3 hours a day blew past 128,000 impressions and 80+ thoughtful comments. To this day, my DMs are flooded with the exact same question:
“You talk about making utilization visible. But how do you actually do it?”
Today I’m pulling back the curtain on our core technology.
Historically, labs had only 3 terrible options to track instrument usage:
- Manual paper logs: Inaccurate, hated by scientists, and always out of date
- Native instrument API integration: Expensive, slow, brand-locked, and impossible for legacy equipment
- Cameras: Massive privacy concerns, and they can’t tell if an instrument is actually running vs. just sitting there with someone standing next to it
We spent 3 years building the 4th, better way: AI-powered power signature fingerprinting for non-intrusive utilization monitoring.
The principle is simple: Every electrically powered instrument generates a unique, unmistakeable power curve when it runs, idles, preheats, calibrates, or faults — just like a human fingerprint.
We’ve trained 100+ specialized AI models on every major lab instrument type: GC/MS, LC/MS, NMR, centrifuges, incubators, HPLCs, and more. You simply plug our smart socket between the instrument and the wall. In 30 seconds, the system auto-identifies the instrument, matches the optimal AI model, and starts streaming real-time, accurate utilization data.
Most importantly: Our system was built by lab people, for lab people.
It does not count preheating, sample prep, method development, or scheduled maintenance as “idle time.” It also does not count an instrument left on overnight with no samples running as “active time.” You can fully customize every work state to match your lab’s unique workflows, so your utilization metrics reflect real usable capacity — not just a meaningless uptime percentage.
Here’s what that looks like in practice: A top 20 global pharma client self-reported their instrument utilization at 55%. After 2 weeks of our monitoring, the true number was 28%. 3 months later, by tying this real-time data directly to their booking system, they lifted effective utilization to 62% — and delayed a $1.2M dual mass spec purchase entirely.
This is the power of good data. You cannot optimize what you cannot measure, and we made measurement trivial.