Five Questions with Primal CEO Yvan Couture
Primal is an artificial intelligence company based in Kitchener, Ont, and a member of the Council of Canadian Innovators. In addition to operating at the cutting edge of AI, Primal is interesting as one of the few Canadian firms set up partly as an employee ownership trust.
In the 2022 federal budget, the government announced plans to introduce a dedicated vehicle to make employee ownership trusts easier to create. CCI President Benjamin Bergen called up Primal CEO Yvan Couture to talk about his business, and what employee ownership does for a company.
This transcript has been edited for length and clarity.
Benjamin Bergen: So, for starters, tell me a bit about Primal. How’d you get started, and where’s the company at today?
Yvan Couture: This is a project going back to 2005 that was spun out of another company I was an investor in. So the idea was basically that, as content and data kind of grow, and grow, and grow, it would become more and more difficult to mine it, and that the approaches that academics seemed to be going towards were probably going to fail. And I think that ChatGPT has recently confirmed the thesis we had when we started.
The ability to work with unstructured content and sparse data environments was really the project we took on. It’s taken a lot longer than we had ever anticipated, only because it’s a really, really hard problem.
We ended up doing a ton of foundational research well in advance of trying to do anything from an applied research perspective. At one point, we had 35 full-time researchers, and we spent a lot of time and effort and money on IP.
It took longer, but it’s way bigger than we anticipated. We developed innovations in the knowledge engineering world, by combining knowledge graphs with rules and distributional patterns of language. And today we have the only machine in the world that can create knowledge graphs in real time from very little input. These graphs are contextualized, so they’re relevant to the user’s specific interests, which is all very transformative. And it’s a foundational technology now that can work in a whole bunch of different areas.
When you think about Google and Apple and others, they’ve spent billions of dollars building their own massive knowledge graphs that get used then to deliver stuff. So we started building this machine that can basically take any input — could be a tweet, a comment, an article online — and in a fraction of a second, we create a small, relevant graph that’s a semantic representation of the input. And then that graph can be used for a variety of things: content recommendations, search, semantic annotation, data augmentation, anywhere where these graphs can help. This approach addresses what is called the long tail, and how a user’s interests can be very specific and difficult to model — difficult for both the large language models that have emerged, as well as the large, static knowledge graphs that Google and Apple have built.
And today we have 156 patents in this area. We’ve got another eight or nine actually, that are being filed now in terms of how what we do intersects with larger language models like the one powering ChatGPT.
BB: Really interesting. I am going to change gears just a little bit, though. I wanted to ask you about something else that makes Primal unique — the structure of the company. You guys are set up as an employee ownership trust. Can I get you to talk about that, and what that means for the business?
YC: We’re not we’re not entirely owned by employees. We have a trust that owns shares on behalf of the employees, but it’s a minority position. So we’re not what you would call a purely employee-owned entity.
In our case it’s very simple. You have a trust entity that’s created, and it holds assets on behalf of the trust beneficiaries — in this case, the beneficiaries are the employees of the organization. And it holds actual shares in the company — not options or warrants. There are some detailed legal mechanics on making it work, but on a basic level, it’s that simple.
BB: That’s helpful framing. Has that altered the way that you’ve been able to attract and retain talent in the organization?
YC: Yeah, you know, that’s an interesting question. When we first introduced this at a previous company where I was executive vice-president, we were hiring fast. We ended up going from 100 to 400 people in less than a year. And when you’re a tech company hiring and growing fast, there’s this fantasy land of stock options, right? We would hear about these billionaires or millionaires out of California who made their money from their options.
But the reality is, in most cases, the average person doesn’t make any money off of stock options from a startup. It’s different if you’re a big public company, but from the perspective of using options in a startup, to compensate for lack of paying a fair salary, I think it’s a problem, because in the end, it typically doesn’t work. We’ve always felt that you pay people fairly for the journey, which is a salary and profit sharing. But also, if there’s an event — a destination — they should share in that as well.
We found it to be actually very positive, particularly when the market has crashed recently, and everybody and their brother at these big companies are sitting on options that are so far underwater, that it’s creating a disincentive. Whereas an employee ownership trust bypasses a lot of that. You don’t have to think about a strike price, one hundred percent of the value goes to the employees, and there’s a huge tax benefit.
Ultimately, people want to just be challenged and work in an environment where there’s some autonomy and some respect. Money’s great, but it’s not what’s going to make somebody jump over from another company.
BB: That’s a really helpful way to frame it. One of the things that actually brought CCI together as an organization was real concerns tech CEOs had around stock options, and a proposal from the government to change taxation in a way that would make them less useful for tech companies and workers.
One of the things that we saw in the 2022 budget, was that they’d make employee ownership trusts easier to implement. What’s your feeling about that?
YC: Personally, I think there’s nothing for them to do other than to create some sort of memo that acknowledges employee ownership trusts and says these tools are there. It’s really no different than in a family trust. That’s how we built the first one.
I think what the government needs to do is to frame it from the perspective of what it means for employees. I think there’s not a lot of work to do there for the government, I think it’s just a matter of saying, you know, we acknowledge that these tools exist and can be applied in this environment. This is what we qualify as employees. These are the rules of engagement. Just having a clear guide would be helpful and then every accounting firm in the world can say, ‘Okay, we understand what this is, and now we’re going to implement this.’
BB: I think that does a good job of laying it out at the individual firm level. What’s the big picture like? How do you think the Canadian tech ecosystem would be different if employee ownership trusts were a clearer and more commonplace corporate structure?
YC: I think we’d see a greater distribution of wealth creation.
I’m a capitalist, but I also believe that that there’s a fairer way of sharing wealth. And I think for too long, the tech industry has used options to hoodwink people into taking lower salaries.
It’s okay if people want to take a chance on a startup — and sometimes it works, sometimes it doesn’t. But if we want to share the wealth in a more equitable way, employee ownership trusts are a nice clean way of doing it. And there’s some great tax advantages.
I don’t think we have a talent problem in Canada, I think we have a talent allocation problem. So how do we level the playing field on that count? I think giving employees a real stake in the Canadian companies they work for is one way. We’re not going to stop big companies from coming into Canada, and we have to compete with them. I’m okay with that. But let’s level the playing field a bit, and I think one example of how to do that is by using employee ownership trusts instead of the flawed stock options model.
If the federal government’s happy with foreign tech companies setting up branch plants in Canada, then there’s not much we can do. But if we took a more holistic view, and said we want to establish this country as a player in the global innovation economy, what do we need to do? We need to start at the foundational level. It starts with people and starts with benefiting from what they produce. When you think about the work product of an employee that goes to Google or to Microsoft or Nokia, the value of that they create far exceeds the economic value of their salary into our economy when you also take into account things like source code, intellectual property, products and services, revenues, profits, etc.
And that’s where I think we’ve lost. I think where it comes down to, honestly, is that the federal government, actually most governments measure success purely by employment numbers, right? But if we start thinking in terms of economic wealth for Canada, then we should start looking at some of these other structures that better capture the wealth for Canadians.
The Council of Canadian Innovators is a national business council of more than 150 scale-up technology companies headquartered in Canada. Our members are job-creators, philanthropists and leading commercialization experts in the 21st century digital economy.