Elevate your enterprise data technology and strategy at Transform 2021.
It’s no exaggeration to call Element AI one of the top startups in the world right now. With the help of deep learning pioneer Yoshua Bengio, the company is making AI-powered products for the enterprise. And from its beginning in October 2016, Element AI has broken the rules of what to expect from a startup.
Element AI has yet to release a single publicly available product, but the company is already working with customers, has opened offices in Singapore, South Korea, Toronto, and London, and already plays an advisory role to startups that receive funding from the Global AI Fund in South Korea.
At the creative tech conference C2 in Montreal last month, VentureBeat sat down with CEO Jean-François Gagné to talk about challenges enterprise customers face in implementing AI, his company’s first publicly available products, and why he believes AI is allowing startups to challenge incumbent businesses in tech and finance.
This interview has been edited for brevity and clarity.
VentureBeat: Are there any specific kinds of challenges companies encounter in terms of implementation of AI? When you come in the door of a business, what’s stopping them from implementing AI?
Gagné: We’re still all trying to figure out how the IP, data access, and learning component of the technology is going to play out. The market is currently trying to figure that out. The big dynamic that we see is all the conversations about “What piece of IP will your AI keep?”
Because we totally understand that large chunks of the IP we build on top of the data of our customers is their own, but there is stuff that needs to flow back to our platform so we make the product better. And defining that has been something where we need to spend a lot of time every time educating the customers, making sure they see it’s transparent and understand what are they going to own, what are we going to own, and this is one thing that is right now a big puzzle for the industry to solve as a whole. And because we’re one of the first to really do that at scale, I think we’re opening the way there, and that’s one thing that comes to my mind.
VentureBeat: Lately I’ve heard a fair deal of companies talk about how they want to build common sense and perception into AI; it just keeps coming up as part of some evolution to go beyond narrow applications of artificial intelligence. Where do you think we’re going on that?
Gagné: So here’s the thing: Google doesn’t want to interact with the customer. They want their system to run by their own, so the way they’re approaching the problem is by wanting the assistant to have some sort of common sense and figure things out by themselves, and we have a very different opinion there. We believe in human-in-the-loop systems that are highly efficient, highly flexible, highly agile, but where people are still driving and are in control of the governance, still involved in the decision process.
You look at the work they’re doing with DeepMind and everything, where things are just going to take off by themselves eventually — and first, I don’t believe that this is going to happen anytime soon. The black box effect this creates, going down that road and all the potential downside of having people think that they don’t need to look at what’s going on, is extremely bad given the state of the industry, so I don’t think it’s the right way to go at it.
VentureBeat: They say they want to be more transparent.
Gagné: They don’t want to be evil, they want to be transparent … what matters are the actions. What are people really doing? We have to actually look at that concretely, so the way we’re going at this is really to make sure that we maximize explainability [and] transparency as we’re deploying this, to enable our customers and people who are using our tools to have the right governance on top. So [we] invest tons of money in all the monitoring systems and explainability of the models. There’s more effort put in this than in the models themselves, generally speaking, in whatever product we do, whatever thing we roll out, and so of course they don’t want that. They’re at the top already, so they want to make sure that they maintain that edge, so that’s why you’re hearing that from them.
VentureBeat: So then what should people expect from Element AI in the next six months?
Gagné: At this point it’s all about scale, like we’re literally in the bottom of the hockey stick looking to be 500 people before the end of the year, 1,000 next year, so we’re really starting to hit that phase. And then for us it’s all about repeatable deployments: successful, repeatable deployments. It’s just about repeatability, getting our cybersecurity product out there, our product for insurance, banking, and we’ve got stuff coming for logistics companies.
VentureBeat: You guys are in a bit of a different position than most companies, but do you feel like the monopoly that large companies have stifles the innovation that could be happening in the AI startup ecosystem?
Gagné: No longer, actually. I think they’re getting challenged to the point that if they don’t do something about it, they know their business is going to start to — their margins and everything are going to start to diminish, and in the financial sector it’s definitely there.
They’re all chipping away at what used to be very interesting pockets of profit for these financial institutions. And how many hedge funds have you heard, “Yeah, we’re like an AI hedge fund trading commodities on this and that”? This is chipping and extracting a lot of value out of the financial market [that] the big players with their traditional strategy are no longer getting, and on and on.
Like, they’re not afraid someone’s going to come in and go for the jugular or disrupt the fundamental core business, but a lot of the adjacent services they’ve built up over time that were extremely lucrative because their anchor business is so mature, low margin, are being taken away by startups, by the market shifting, and that’s a huge pressure for them to do something.
On the logistics-supply chain side, nobody can deny the impact of AI when it comes to self-driving trucks, automation in the warehouse, the impact of Amazon and all, that’s pushing the market really hard.
When it gets to the media and entertainment information sector, [it’s been] disrupted pretty bad in the last 10 years. And for a lot of these companies that are still there or the new entrant, they need to use one shape or form of highly productive systems, and AI kind of does that.
VentureBeat: The Global AI Fund in South Korea, is that the beginning of something global?
Gagné: So we’ve got an office in Seoul, and that’s different than the fund. The fund’s based in Delaware in the U.S.; it’s a $45 million U.S.-based fund.
VentureBeat: Right, and you’re consulting Hyundai and the other companies making investments?
Gagné: Exactly. We’re a GP in the fund, they’re LPs, and we’re all contributing some tangible assets beyond money. So it’s access to our researchers and helping the businesses we invest in improve their models. The fund is — yes, it’s a test. The one thing that we saw is that our fundamental lab is producing some of the best research in robotics and mobility and certain domains where Element AI is not building solutions.
And we wanted to leverage these insights and kind of get others to benefit from all the IP we were building there, and we needed kind of a vehicle beyond just doing partnerships to create a channel for us to funnel the IP.
VentureBeat: Are you going to be doing more similar kinds of investments?
Gagné: We’ve been approached about that, to do that. We want to see how this one plays out. So the fund was actually finalized in March this year, so we’re starting to look at the pipeline of investment; we’re still kind of early. Our goal is to invest between $3-5 million where we follow, so it’s like 10 to 15 shots we have. So we’re going to take a little bit of our time and make sure that we get it right and see how it goes.
VentureBeat: I know you opened offices in London a couple months ago. What’s the distinction you draw between the work they do there and the work being done here in Montreal?
Gagné: They’re focusing on projects for AI for good, providing access to the tech we have and the advances that we’re making on other fronts to NGOs [non-governmental organizations], and that’s mainly how we’re articulating this team. Eventually of course our European office will be a full-stack office. The way to think about is like Singapore is our main Asia office; we have customer success, customer delivery, project management, and sales. Product, platforms, research [are] all conducted in Canada but mostly Montreal, 240 people in Montreal and 60 in Toronto; Seoul, for now, sales and project management — they support customer support success and delivery for Singapore.
London for us is an incredible talent pool that we want to tap into and research relationships. We always wanted in the plan to make sure that as we’re building the tech, we provide access to everyone — it’s always been part of the vision for the company — so then dedicating the office as we’re growing it to be responsible to build products and projects for good made a lot of sense, and that’s the rationale.
VentureBeat: I was listening to Andrew Ng talk recently about some industry verticals being better suited for applications of AI than others because they just had more labeled digital data. Are there any specific industry services that …
Gagné: Financial services — they’re digital enterprise already, and they’ve tried to do something for a while now with IBM and Palantir and they didn’t get that far. Now they’re more mature: They failed, they learned why in general it’s a market that’s absolutely ready, and, I mean, you would look at our pipeline and it shows.
Gagné: Yes, there’s many others. So financial services is one, and the other is the supply chain. And we think that it’s going to take more time for supply chain, but it’s going to be so transformative and impactful for society, that’s really where we want to invest our brain cycles. These are the two sectors we’re focusing on.
In terms of readiness, I think media is more ready; all the consumer tech is obviously ready, but the problem is it’s a bit cornered by some players — in China it’s Tencent, then Facebook and Google and Amazon. They’ve cornered a little bit the consumer market. There’s still a lot of opportunity. I think the media and entertainment industry has huge opportunities; digital businesses have lots of table data.
VentureBeat staff writer Khari Johnson participated in a panel discussion about AI for good at C2 cosponsored by C2 and Element AI. As a result, Element AI paid travel and lodging expenses for attendance at the conference. No Element AI employees took part in the panel discussion.
VentureBeatVentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the subjects of interest to you
- our newsletters
- gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
- networking features, and more