Toronto’s Canvass Analytics strikes big on Google investment, eyes global partnerships
The investment is Gradient Ventures first introduction into industrial A.I driven analytics and predictive maintenance for IIoT solutions.
For Humera Malik, the path for co-founding and launching the Toronto-based Canvass Analytics traces back to 2005 where she worked for Bell telecommunications as an Assistant Director working on the implementation of smart buildings.
“On the telecoms side, where I started, we were adopting a lot the smart solutions and I got to help build smart solutions with things like smart metering,” Malik says. “It wasn’t exactly on the industrial side, but it was in the smart utilities that I really was able to learn the genesis of the world I’ve found myself in now.”
Focused on industrial implementation of predictive analytics, Canvass Analytics made headlines earlier this year with a five-million-dollar investment from Gradient Ventures, Google’s A.I focused subsidiary. Interesting as a success story for Canadian startups, the news is noteworthy since it’s the first investment Google has made into the A.I driven industrial analytics market.
At it’s core Canvass Analytics tackles the “Data rich, information poor,” situation many companies find themselves when implementing IIoT solutions. Working primarily with manufacturing companies, Canvass skews away from big data projects that don’t lead to actionable results. Instead focusing on predictive analytics and scalability that’s useful to customers immediately.
That speed and scaleability begins by trying to answer three questions about the data clients collected, allowing Canvass to best serve their needs:
- How can I gain a competitive advantage from this data
- How do i meet the growing demands of my market using this data?
- How do I optimize production so that I’m prepared for the future?
“Who better to partner with than Google? who’s actually building the future of A.I,” Malik says. “We’re really proud to be their first investment into industrial analytics and for us it’s all about ‘Where can we extract learning opportunities from? And who are the people that are setting these technology standards and creating these platforms?’ Because these are not solutions, solutions are built by people like us,” she says.
From shifting data to streamlined automation
It’s using analytics to help automate certain production line tendencies. For example, in real-time, Canvass’ A.I algorithms can watch product quality as it’s being worked on, providing feedback to an operator. If anything were to fall outside of the product parameters set by the operator, the A.I would let them know before the quality of the product was jeopardized. Using the automotive industry as an example, Canvass is able to use their AI-powered analytics platform to predict failures and optimize assets to the tune of 10 per cent reduced production downtime.
Another area were their platform is working in regards to optimization, is increasing productivity in energy companies. In one case, Canvass Analytics is helping a global agriculture processing plant optimize their co-generation turbines which has led to a 13% improvement in fuel efficiency while reducing CO2 emissions by 10+ million pounds per year.
This speed and scalability of their solutions are two of the reasons Gradient Ventures decided to partner with Canvass Analytics. This flexibility is generated in part through microstructure architecture—a variant of the service-oriented architecture—which are lightweight, easy to develop and offer repeatability and consistency. They’ve adopted this strategy over time by asking themselves ‘how do we allow our customers to do more with what they already have?’
In a statement released in conjunction with the investment, Gradient Ventures founding partner Ankit Jain said “Autonomous operation is the holy grail of manufacturing and AI is the game-changer that is making it a reality across the industrial landscape. We’re backing Canvass Analytics because of its unique approach to implementing AI and predictive analytics quickly and in an automated manner, without the need for lengthy and often cost-prohibitive consulting engagements.”
Microstructure architecture 101
If you think of microstructure architecture as nimble, fast and responsive — a monolith, or “giant,” application is the exact opposite. Whenever changes need to be made to a monolith application, it takes time and can be expensive. It means potentially deploying all new software over a few new lines of code, or scaling an entire operation when you just wanted to scale a few specific functions. The companies mentioned above shifted to a microstructure architecture over the years as their customer base demanded a more versatile and fast response.
With a microstructure architecture, each function is individualized, separate from the whole while being very easy to work with and make changes to — think modular analytics. This type of system is useful for times when you can’t completely predict the types of devices or functions that one day might be needed to run the application. The small decentralized nature of these systems allow programmers to revamp and create new solutions quickly. Of course. it’s not a perfect strategy: testing can become complicated and tedious due to decentralized deployment and the increased number of services can result in information barriers, but it’s a strategy growing in popularity. Amazon, Netflix, Twitter, eBay and Paypal all work off a microstructure system, but this wasn’t always the case. They were once slow moving architectures that responded to the growing demands of their consumer base.
Security and the future
With the explosion of IIoT over the last few years, security concerns and proper implementation are on everyone’s mind, include Canvass Analytics. They recently announced a partnership with Microsoft’s Azure platform, and deal with third-party software security groups when it comes to things like cross-border sharing of data for example.
“We rely on partners to provide the different layers when it comes to the compliance and security side, because we deal with some of the most sensitive data that you can think of,” Malik says. “Some of these things become challenging when you’re dealing with an open consumer service and some people might not care, but for us the most important thing is our customer data.”
The next 18 months for Canvass analytics are focused on expansion and further partnership. Additionally, the company is thinking global, currently working in Europe and Asia. According to Malick the company is planning to move into “mature and progressive markets where industrial IoT is being implemented, even at the government level.”
On the partnership front, specifics are obviously still under wraps, but Malik did note that “huge partnerships,” with companies who already exist in the industrial space from a back-end, analogue perspective are forthcoming. For Canvass, it’s about “how do we integrate ourselves with them to become this intelligent, industrial analytics platform that enables them to go out and create the stickiness in the market.”