AI meets heavy industry: Why we invested in Senseye
Modern manufacturing facilities are a technological marvel. It’s impressive (and a little humbling) touring around a state-of-the-art factory with a high degree of automation. Robot arms, smart conveyors, hundreds of machines moving in harmony, computer-controlled and responding to inputs from an array of sensors.
Except when they stop moving. Unplanned downtime is a real productivity killer, especially for ‘just-in-time’ operations with small profit margins. Think of a car assembly plant pumping out a new vehicle every minute — if a critical machine fails for an hour or two, that can cost millions of dollars in lost production and other costs further up the supply chain.
We believe that predictive maintenance will be an important trend in manufacturing over the next decade. Today, 90% of maintenance work is still reactive, often in the middle of a crisis. With more sensors deployed and better availability of data, new diagnostic and prognostic (‘remaining useful life’) capabilities are becoming possible, providing advance warning of an impending failure or degradation. This allows maintenance teams to stay one step ahead, avoiding those expensive scrambles to fix or replace a broken part.
That’s why we’ve backed Senseye, a leader in predictive condition monitoring and prognostics, in their £3.5m Series A (announced here). Having proven that they can make dramatic reductions in downtime in some of the world’s largest manufacturing facilities, they are on a mission to bring predictive analytics into the mainstream. They’re one of the highest scorers against our AI assessment framework, out of more than 250 UK AI companies we have met.
Senseye has developed a unique series of machine learning algorithms that take inputs from sensors — movement, vibration, temperature, current etc.—on each device on the factory line. Their platform builds models of baseline performance and failure modes, and detects small deviations that can be leading indicators of a problem.
But smart data analytics isn’t enough. To succeed in this space requires a deep understanding of the engineering and the factory context. Any successful solution has to consider how the maintenance team work, understand the potential bottlenecks in the line, and differentiate between an impending failure and a change of mode, or some planned maintenance that the team forgot to log.
Senseye’s team has all of this in spades. Co-founder and CTO Robert Russell has spent the past 20 years designing and deploying custom condition monitoring and prognostics solutions for critical machines like aircraft engines. Working in this field with co-founders Dr. Simon Kampa and Alex Hill over many years, the three teamed up with machine learning expert Dr. Daniel Read in 2015 to combine the best parts of the traditional approach with cutting-edge data science. They have created a product that sorts the signals from the noise, surfacing only actionable insights.
Their customers include leading automotive and other manufacturing conglomerates, in the UK, Europe and Asia. It works in the real-world environment of the shop floor, and the maintenance teams love it because it makes them more effective and their jobs less repetitive.
You could say they are selling ‘uptime-as-a-service’.
VCs tend to talk about software and typically don’t give much thought to the multi-trillion-dollar global manufacturing economy that produces physical products (unless they are drones or chips for mining bitcoin). As we at MMC Ventures explore how artificial intelligence is changing different parts of our economy — see our in-depth report, the State of AI 2017 — we believe that many of the highest-impact applications will be in the industrial world.
This isn’t our first investment at the intersection of software, new technologies and heavy industry — Sky-Futures is changing industrial inspections using drones and analytics, and Breathing Buildings (which we successfully exited last year) pioneered new energy-efficient cooling technologies. Don’t expect it to be our last!
Click here for the full press release, or visit mmcventures.com
