- PreSeed Now
- Posts
- A.I. to give machines a longer, smarter life
A.I. to give machines a longer, smarter life
Manulytica wants to help the industrial sector avoid costly breakdowns
PreSeed Now brings you a different B2B or deep tech early-stage startup every Tuesday and Thursday - in depth.
Subscribe for free to get it straight to your inbox:
Can you make a dumb machine smart? Today’s startup believes you can. Below, we’ll tell you all about Manulytica.
But first…
Things you should know today:
Good news for angel rounds from the Chancellor’s budget yesterday: the controversial rules that cut the number of people who could be considered high-net-worths are set to be reversed. Well done to everyone who campaigned on this issue.
Smartphone security startup Nuke From Orbit has raised a £500,000 pre-seed round. Investors include Oliver Bridgen, co-founder & COO of Ballinger Group.
– Martin
Manulytica wants to give industrial machines a longer, smarter life
In summary:
Premium subscribers get the full version of this article, plus a TLDR summary right here, and access to our Startup Tracker for updates about what this startup does next.
Industrial machines can often be used for decades. They’re workhorses, reliably churning away… until they break and seriously disrupt the businesses that depend on them.
Manulytica is a startup that wants to bring modern day smarts to machines that were built without any such intelligence built in.
Its hardware is designed to detect when a fault is going to develop using on-board A.I., so a machine can be fixed before it stops working.
“There is enough information on manufacturing equipment to be able to tell you when and where it will fail, where and when you will have quality issues, and where and when you can improve what you're currently doing,” says founder William Fish.
“So Manulytica has been set up to help manufacturers unlock that hidden data, to be able to understand how things are performing, and how to improve.
“We're then trying to implement machine learning and artificial intelligence, not at the cloud level, not at the edge, but actually on the machines themselves so they can take advantage of making things actually smart.”
How it works
Because the data about how machines are running can be proprietary and difficult to access, Manulytica has developed a device that connects directly to a machine.
“It allows you to connect directly to the sensors or the signals being sent to actuators, so you can find out what's actually being done as it happens,” says Fish.
While developing the device, they realised that cabling for such a device can be expensive, and so the hardware is wireless.
The hardware monitors the running of the machine, identifying when it deviates from normal running.
Manulytica’s current prototype, which the startup says is approaching commercial readiness
Fish says Manulytica’s on-board A.I. makes use of a neural network that learns about how the machine operates. But while some predictive maintenance models will simply say ‘your machine is going to fail around 6pm on Friday’, Manulytica’s hardware goes a step further.
Let’s say a motor is likely to fail, the startup’s hardware can send signals to the machine to reduce the motor’s speed.
“In that scenario, it could change it so your motor can’t run at 100% speed. It goes down to a maximum of 75%. And if you still haven't done anything, drop it further. It's kind of like ‘limp mode’ for your factory, instead of your factory completely breaking down and producing nothing and leaving you stranded,” says Fish.
You might assume that the next stage would be for the device to send the data it collects to some sort of software dashboard. But Fish says they’ve deliberately avoided this.
The use cases and priorities of customers can be so diverse that it doesn’t make sense to offer a ‘one size fits all’ approach to displaying the data.
Instead, data from the Manulytica device can be analysed in whatever way suits the customer, and it can be fed into existing software they might already have.
“We're not trying to do a beautiful interface… There are lots of really good manufacturing execution systems out there. There are lots of very good ERP systems out there. I don't want to reinvent the wheel,” says Fish.
William Fish
The story so far
Fish grew up in Silicon Valley as the son of an engineer and a musician. After beginning a career in music, he shifted into the manufacturing sector.
The seed of the idea for Manulytica came from a project he was involved in at the turn of the century to install sensors on printing presses.
“It was very simple. Grab the data, put it into a database, and then we put a front-end over the top. I was heavily involved in this and thought it was fairly straightforward. I thought ‘Why isn't everybody doing this?’,” says Fish.
“So I learned many moons ago, something that I repeat a lot, which is I want to be able to be at a bar on the beach in Barbados, and know what's going on, as long as I have an internet connection.
“If you run a factory, if you run in a warehouse, why can't you know exactly what's going on from anywhere in the world? And that's what we're trying to do here.”
Fish designed and built the first prototype Manulytica device during the Covid lockdowns when he had a lot of time on his hands. While the device is still at pre-production stage, it’s now on its twenty-first iteration.
The device is currently being refined to prepare it for commercial launch. Fish says he’s in discussions with manufacturers to launch pilots as the startup “inches closer” to launching a product onto the market.
And there’s more!
Premium subscribers get the full story about Manulytica:
Upgrade your subscription now to learn about:
Manulytica’s funding and investment plans
Founder William Fish’s vision for the ultimate goal of the startup
How Manulytica compares to the competition
What challenges face the startup as it grows
Subscribe to Premium Membership to read the rest.
Become a paying subscriber of Premium Membership to get access to this post and other subscriber-only content.
Already a paying subscriber? Sign In