AIP Podcast EP 70 – Industrial AI Knowledge Engine in Action by Canvass AI

This episode’s guest, Humera Malik, Co-Founder and CEO of Canvass AI, joins host Anne to explore how AI is reshaping the future of industrial operations. Humera shares her journey of pioneering AI in conservative sectors like oil & gas, manufacturing, and food production—long before AI became a buzzword. She discusses the biggest challenges facing industrial transformation, from change management to overcoming skepticism about AI’s role in augmenting rather than replacing the workforce. Humera also explains how Canvass helps operators cut energy and material costs, navigate data overload, and empower engineers to achieve “10x” productivity through actionable insights. Backed by global leaders such as Alphabet and Yamaha Motor Ventures, Canvass AI is driving sustainable impact across industries while preparing the next generation of operators for the AI-powered future. 

Follow AIP Affiliate, Canvass AI
Website: https://www.canvass.io/
LinkedIn: https://www.linkedin.com/company/canvassai/
YouTube: http://www.youtube.com/@canvassai1708

Follow AI Partnerships
Website: https://www.aipartnershipscorp.com/
LinkedIn: https://www.linkedin.com/company/aipartnershipscorp/
X: https://twitter.com/AIPartnerships

The AIP Podcast is hosted by Anne Cheng, on behalf of the AI Partnerships, a Railtown company

Transcript

SPEAKER_00 0:00

The fifth industrial revolution is underway and artificial intelligence is growing still much of it. And yet, when you get deep into the industrial transformation, many operators and engineers still struggle. Welcome back to Yeah and the Delphine of the AIP project. And I am entertaining your followers, and today I'm speaking with Kumara Mannek of Candidate AI. Now candidates AI is unvention to democratize AI across enterprises by making it accessible, usable, and most importantly, impactful for every single employee that try to empower organizations with tailored AI solutions that enhance decision making, optimize workflows, and drive innovation in every corner of the business with a focus deadset on some of the world's fundamental industries like oil and gas, healthcare, and manufacturing. And Humera has been recognized by women in AI and is passionate about applying AI and artificial intelligence to optimize industrial operations, augment workforces, and achieve sustainable impact. Humara, thank you for being on our show today.

SPEAKER_01 1:14

And thank you for having me. Really excited to be here and be part of this network that's bringing about this change in the industrial side.

SPEAKER_00 1:23

Congratulations, first and foremost, for being backed by such notable names like Alphabet and Yamaha Motor Ventures. And your backstory particularly is fascinating. Take us through your journey.

SPEAKER_01 1:38

So I always I would, in these kinds of contexts, I would start by saying, you know, AI wasn't necessarily, you know, or Canvas AI didn't necessarily come into existence post this revolution of AI that we have seen in the past couple of years. We've been at this when AI was the scary word that nobody wanted to actually acknowledge or or even adopt. So back in when we started, AI really was something which was uh which was not the the the the term or the technology that anybody was comfortable with. But the reason you see people like Alphabet and people like Yamaha backing us is when when I was working on this, which was really I was working in the industrial sector and constantly seeing people just hoard a lot of data. And to me, that was a real problem. And so when I went in to solve that problem, AI became the necessary technology. And it wasn't necessarily the hype, it was the necessary technology when you're dealing with behemoth of data, when you're dealing with large volume, verosity, and frequency of data, AI is the natural technology that you adopt. So I had to go and crash course and learn all about AI at that time. And that's what I came up with is that was the really the best way to go out and solve problems in these industries. And Google came along and then Yamaha came along in Alphabet now. They came along because I think at that time it wasn't necessarily the thing that was well recognized and known. But obviously, companies like this that are that are themselves in in um in innovation, that's what they back is innovation. And when we went and tried to do this in the industrial sector, there weren't many people trying to do and apply AI in industrial processes. One is you're looking at industries, you look at energy market and especially oil and gas, and you look at food and other steel industries. These are very conservative industries. And, you know, players that have existed, they've existed for decades and decades. So for somebody, a small startup coming out of Canada and going out and trying to actually transform an industry wasn't the easiest journey, but I would say it was because of this kind of a backing with people like this that that had the belief in it, that we went out and actually then started to get into some of the some of the players that really weren't necessarily ready for it, but they at least were open to bringing it in and seeing what this could do. So the art of the possible. And so that's really what Canvas has been doing, with obviously the backing of Alphabet, Yamaha, Real Ventures, and others, and obviously Canadian government as well. We've been able to go out and really, I would say, is make at least our mark and a dent into this to say we were one of the pioneers that brought about AI into an industry that necessarily wasn't very um open and adaptable towards bringing new technologies into their operations.

SPEAKER_00 4:47

That's quite the pedigree you've got there. But you know, when you moved into the industrial transformation space, um let's talk through some of those challenges that you faced. Um, what were the key challenges that you faced in the industrial transformation space?

SPEAKER_01 5:04

So I think the people would think that these are very old industries. So probably the biggest challenge should be data. But I think, and surprisingly, data wasn't the biggest challenge that we faced in these industries. The biggest challenge that we faced really was change management and the mindset. AI, even today, is looked at it in a way where it's going to replace jobs, but people are rushing to adopt it because they're looking at it as a way of I can actually have a have a path towards not just saving my job, but actually some newer jobs that I could create with it. At that time, it wasn't so. It was really where if you go in with AI, the first thing you would hear from engineers and operators would be, no, we don't need this, we know everything. Does AI, I've actually I've actually gone in into places where even the senior executives told me, so do you actually know this process? It was it was in an automotive environment. So do you actually know the welding process? How long have you worked in welding? And those kind of pushbacks that you would that we saw at that time, which was if you're coming in, you better be an expert. Do you have a certification or a degree in that? And that's the only way we would let you enter into these places. So it wasn't necessarily we don't have the data, it's you do you have the credibility to touch my equipment? Do you have the credibility to come in and touch my processes? That was the biggest barrier that we faced. Because AI is coming in and learning, it's not necessarily we're throwing humans at it, and when you throw AI at it, they they panic because that's not what they're used to. So, anyways, that I think the the difference between other industries versus this is that you have a you have no shortage of data. What you have shortage of is is basically the adoption of change. And so to me, change management has been and continues to be one of the biggest challenges that we face going into these industries.

SPEAKER_00 7:00

That's crazy, but yeah, you're absolutely right. And talking about industrial IoT particularly, a lot of IoT devices today collect loads data, but a lot of them don't collect a lot of insight. In fact, collecting such amounts of data can be difficult given that it requires a lot of significant compute and strain on connectivity resources, uplink and downlink. So, how do you help your customers navigate the issue of cutting through this noise?

SPEAKER_01 7:32

So you're very right, Anne. There is definitely everybody is in the race to go out and collect data. This has been going on and more so now, even before, right? Because everybody's going out and trying to collect data because there's a certain belief that exists, not just in the industrial sector, but overall, that the more data that you have, the more power that you have. And to me, the more data that you have, the more noise that you have created now. And I think the biggest one of the learnings that came for as we were going out and implementing this for our customers was it's not about the volume of data, it's about having the right data. And then not necessarily that you need to go out and join this bandwagon and start to collect all this data and put it all in the cloud because that's not necessary. Sometimes, if you want to create, if you want to control a process, it's better to collect the data right at the site, right at the edge, so that you know you can at least make decisions at the right time at the right place as well. So instead of following, hey, you know, we need to, we all need to go out and implement hundreds and thousands of sensors and and devices and collect data. It's what we have been able to do is work out with our customers to to have them understand that this is not a race towards getting more data, the race is towards how efficient can you be with the the right data, how you can automate your processes, how you can get what we now do is 10x engineer, because there's a real shortage of people on the ground for these kinds of um processes. So, how can you empower an engineer with AI to create 10x the power? And just data doesn't do it, just having loads of data doesn't do it. So when we work with our customers, really what we're looking at is analyzing first is what is the data necessary to make that decision, and then reverse engineer back into what is needed into the cloud, and then what is needed at the edge, and what is needed on the fingertips of what kind of data is needed at the fingertips of these engineers to make decisions. So to me, um the this race that people have taking towards everything needs to be on the cloud and we all need to move to the cloud, put all the data and put more sensors. That to that to me is not the right strategy where we have seen our customers drive ROI from these investments. It's really about having the right data, contextualizing it for the right set of people. Engineers, they need data so they can it they can make decisions on the design, they can make decisions on maintenance. Operators need the data so they can make decisions on smoothly running the process so they can deliver the right product at the end of the day. RD needs it for research purposes so they can continuously feed the engineers and the operators. And so having that data and then contextualizing it to the right people at the right place and so that they have it available really is the strategy that we have been we have been implementing with our customers.

SPEAKER_00 10:34

That's amazing. And that's really what you mean by get driving value out of people's data at speed and at scale. Let's take it down one little notch. If I were a small-scale owner-operator of a food manufacturing facility, how can you help a person like me?

SPEAKER_01 10:55

Well, if you're if you're a owner, operator, manufacturer, a small food facility. So in the food industry, there are two things that really keep you up. One is the cost of energy, and second is the cost of your raw material. Those are the two biggest costs. And food industry is one of the industries that really is surviving on razor thin margins. This is an industry which has the least amount of margins, and so they're tracking everything. So basically, today you're tracking most of the things through spreadsheets and some automation where you might have put some sensors so that you're collecting the data. And so, if you're coming in or I'm coming in and advising you on any of this, is the first thing is is what are the controls that you have in place, and how do we help so that you can either predict or recognize patterns so that you can reduce the energy costs or you can reduce the raw material cost. So those are the two things we get started with. Energy being, I would say, a much simpler path because most of the people have at least got some kind of energy assets that they are managing. So those could be co-generation assets, they could be running turbines and boilers, or those could be any other energy efficiency equipment that they're energy equipment that they would be running, that you can easily drive efficiencies. Because if you can use AI basically to easily just uh calculate what your production is and how much energy you're using, it can actually tell you then how to optimize it, really. Okay, if you're running the plant at let's say you have to produce 100 ml of something and you were running the plant at like 10 kilowatts, then it can it can what AI can do for you is actually reduce it by 30%. And this we have proven time and time again in these industries, especially in the food industry. Same thing with raw material, it can actually look at the different combinations and it can help the operators control different things. Simple things, controlling temperature at the right level so that you know you can control the moisture so that no waste gets created. Those are very, very simple things which a small food manufacturer can start to implement just on the basis of the spreadsheet data, and they can reduce their energy costs anywhere between 20 to 30 percent.

SPEAKER_00 13:12

That's impressive. Thank you, Homera. Um, well, it's all the time that we have, but one for the road, um, Hermera, what's on your horizon? What are your plans moving forward to propel the future of industrial intelligence?

SPEAKER_01 13:27

Um, I think we're living through very exciting times of AI. Um, as I said, we were early in the game as the pioneers of bringing it into the industry. The the shift that I'm now seeing is there is a more pull rather than a push. And I I really I'm I was waiting for this wave to happen, and it has finally happened. So, really, the future seems, especially for the industrial sector, there is so much opportunity because we have uh an aging workforce, and this is really where AI is coming in and showing its muscle power, being able to help the new wave of engineers and operators that are coming in. I'm super excited because they're actually super excited about uh the use of AI in their day-to-day operations.

SPEAKER_00 14:09

That's brilliant. Homera, I've really enjoyed speaking with you, and I think I've learned a lot from you today. So, to our listeners, thank you once again for being with us. And if you've enjoyed this episode, don't forget to like, share, and follow us, and even send this podcast to someone who could benefit from it, like I have. Once again, my name is Anne. I am your host, and on behalf of the AI Partnerships Corps, and my guest was to marry Monik of Canvas AI. Until next time, stay safe, folks.

SPEAKER_01 14:40

Thank you, Anne.