FounderFuel general manager Sylvain Carle is readying to graduate his fourth cohort of startups since taking over as leader of the Montreal-based accelerator program two years ago.

The general manager told us things are different than they used to be. He was making finishing touches to the program’s Demo Day event scheduled for tomorrow evening.

Most of tomorrow’s graduates have some form of machine learning within their platform, a far cry from the days when mobile apps nearly dominated entire Demo Days.

“When I came in software was eating the world. It was all apps, cloud, marketplaces and social. I wanted more industries represented and I wanted diversity,” he said. “Now we’re seeing artificial intelligence (AI) as this ‘fourth pillar,’ and broadly applicable across all industries. It’s really coming together nicely.”

It was four months ago that Carle and his team at FounderFuel made their call for applications for the new cohort, slightly emphasizing machine learning and AI. Within a few months those two terms would explode in the world of tech startups, as investors scrambled to fund the best and brightest founders proficient in things like “neural networks” and “deep learning.”

“The question we had was could we in 2016 find startups that are applying machine learning to a specific domain, and that’s been proven true. By looking at applications we found a great snapshot of how the current state of AI and machine learning is being applied in a business context. But we’re still really, really early on,” said Carle.

About half of those startups graduating tomorrow are using AI as a core driver of their platform. Carle’s goal was to have women CEOs leading 50 per cent of the companies, but only two of the six companies check that box. The only other hint he gave us was that the cohort includes one team within agricultural technology (agtech) and another in educational technology (edtech), both of which are driven by AI.

What intrigued Carle in selecting the cohort was how the teams were using data. How were they getting data? Were they looking at data sets and seeing insights that others missed? Could they predict outcomes from limited data?

From there it was how these teams were using AI to drive their businesses. They needed to move past ‘level one,’ which was monitoring past data and making conclusions from it. If a team is really good at understanding what happened in the past, they ideally facilitate real-time decision making within the tech and ultimately predictive decision making. “This is what were seeing with ads or financial trading,” said Carle.

Take the agtech startup, he said. The founders needed to get their hardware and their sensors in the hands of growers so they could collect data and their machine learning could understand the conditions for improved crop growing.

Typical of FounderFuel’s Demo Day process, we won’t know that startup’s name until tomorrow at about 4:00 pm when all the teams pitch on-stage.

FounderFuel advocates will argue the program was the first accelerator program in Canada. It was certainly one of the first alongside other vestiges of the past like Toronto’s Extreme Startups, Waterloo’s Hyperdrive and Vancouver’s GrowLab.

FounderFuel’s parent venture capital fund, Real Ventures, is statistically the most active seed investor in all of Canada.

Tomorrow’s cohort will be the program’s ninth. Ian Jeffrey, a cofounder of this blog, was the program’s General Manager for five cohorts before Carle took over. Over five years the program counted 73 teams as having gone through the three-month program, 37 of which went on to raise money and are still in operation today. Four of those 73 went on to raise a series A round of funding. Ten of the companies “exited,” or sold their company to a larger competitor, and 26 completely shut down operations.

Carle typically has to cut ties with one startup per cohort, usually midway through the program. Early validation isn’t happening, perhaps there’s still no product or cofounders are fighting.

But as common accelerator wisdom goes, if just two of those 73 companies that FounderFuel funded turns into a company with real profit and strong growth then the program can be seen within a successful lens.

“So if you look back to those 22 companies out of 73 that raised proper seed rounds or a series A, we would say it’s been working pretty well,” said Carle.

Teams entering the program today are hitting milestones faster, as opposed to two to three years ago. Carle noted Flatbook, which had already booked $1 million in sales by its Demo Day last year and Sharethebus, which is bringing in somewhere around that same amount during the year.

Both companies entered FounderFuel within the last 18 months.

“We have a better level of understanding of what worked and the companies are doing more faster with less money,” said Carle. “The path that used to be five years is two to three years now. Now some companies rocket ship with $300,000 in seed money, whereas it used to require $1 million, $1.5 million.”

Things are also a far cry from five years ago, when accelerators in general were few and far between in Canada. Today at least five other programs exist in Montreal, including longtime programs like TandemLaunch as well as fledglings like InnoCité MTL, Concordia University’s District 3, ÉTS’s Centech, Ecofuel and more.

For Carle, an increased demand among founders means a new reality within Montreal. How FounderFuel will fit in through all the clutter remains a dominant question in his mind. He doesn’t claim to have the answer, but one can bet he’s planning for something.

“Five years ago there were no accelerators in Montreal. That’s what I’m thinking about for the future: what’s the role of FounderFuel in the ecosystem? What’s the next chapter when there’s 10 other accelerators in Montreal?”