TORONTO – The 15 companies that presented at startup-boosting non-profit Next Canada’s inaugural NextAI Venture Day provide an excellent snapshot of how far Canada’s artificial intelligence (AI) ecosystem has progressed in the past year, co-chair Tony Lacavera said.
Each offered investors a viable product. Each was the result of a brand-new program designed to facilitate the creation of an AI-based solution to a specific business-related problem. The young entrepreneurs running each company were all given access to the best mentorship, education, and networking opportunities Next Canada – known best for its Next 36 program for young entrepreneurs, which it’s run since 2011 – had to offer.
And on Sept. 20, each was given the opportunity to present their solutions to investors.
“Six of the 15 companies – each of which are around nine months old – are already generating revenue,” Lacavera said. “Six have paying customers. I think that speaks volumes.”
While Next Canada academic director Ajay Agrawal compared the 2017 version of the program to listening to the Beatles’ first album – “it’s not polished, but there’s a raw energy” – Lacavera he was extremely pleased with the presentations, which were divided into AI-based solutions for human resources, fintech, marketing, and what event host Amber Mac called “potpourri.”
SortSpoke, for example, allows enterprises to automate their data entry operations.
“There are still four million people onshore who spend a lot of their day just reading documents and extracting data from them,” Toronto-based founder Jasper Li told the audience. “It really doesn’t make sense for humans to be doing that work anymore.”
Li speaks from experience: Before founding SortSpoke he served as a management consultant, helping enterprises transform their operations through digital transformation.
“One problem no enterprise could ever solve was how to get rid of the people who are just sitting there reading documents and contracts,” he said.
The key difference between SortSpoke and its competitors, Li said, is his acknowledgement of the reason companies have been reluctant to embrace data-scanning technology, even though it’s been available for years: It’s simply never been as accurate as humans.
“You need to keep humans in the loop,” he said. “Many companies value the expertise of their staff who can read a 30-page contract and know how to find exactly the data that they need. We just want to take their value and expertise and let them be 10 times more productive.”
SortSpoke’s machine learning-based algorithm, Li said, is capable of learning how to extract the data a company wants most from its documents in less than a day, and can then extract, sort, and export data for human experts to review within minutes.
“You’re not going to get rid of all of those four million people,” he said. “You’re just going to get rid of 95 per cent of them… while keeping your human experts.”
CEO Marlon Rodrigues was inspired by Environics research indicating that while 29 per cent of Canadians trust product information delivered by bloggers they follow, a much higher 74 per cent trust recommendations from someone they know.
“Our premise is that trust, not follower count, predicts purchase influence,” he said. “We trust the people we know two and a half times more than social media influencers, and as a company, we want to bring the type of compensation available to social media influencers to the people who actually drive purchase behaviour.”
Users who install the Truffles Chrome extension are automatically notified whether a product page they’re about to share is associated with a cash reward. If it is, the user is sent a message with an alternate link that allows them to collect a reward should the recommended product is purchased.
At the moment, Truffles is in open beta. The company has signed an affiliate deal with Amazon, and to entice new users is donating its entire share of each sale to them whenever a recommended product is purchased, though Rodrigues acknowledges the current arrangement cannot last forever, and that eventually Truffles will retain a portion of the proceeds to fund its business.
For now, though, Rodrigues is enjoying helping build relationships between brands and users who never would have self-identified as a social media influencers.
“Social media influencers have been getting paid for the influence they assert for years now, while the people who are actually driving the majority of purchases by sharing links with friends are getting nothing,” he said. “The technology exists, the paradigm is well understood, yet nothing has been done. So this is us recognizing that gap and using technology to solve it.”
From former Google engineer James Damore to the sweeping changes announced at Uber headquarters in June, the tech veterans attending NextAI Venture Day didn’t need to be reminded of the leading role company culture plays human resources retention – or how difficult it can be to change.
That’s where Cultura comes in. Using a combination of machine learning and the expertise of University of Toronto professor András Tilcsik (co-author of a book on the subject, scheduled to be published next year) CEO Katya Kudashkina has developed a mobile-friendly platform that companies can use to gauge how employees feel about their operations, and suggest improvements.
“[The platform] essentially has two parts,” she explained. “The first is a mobile survey, which assigns ongoing assessments to employees on a weekly basis. All questions are binary – yes/no – and the answers all flow into a dashboard, where [managers] get to see what the problems are, along with analysis of why they’re happening and how to fix them.”
A great deal of research goes into developing both the questions and machine learning used to analyze the results, Kudashkina said, with queries such as “It is important to avoid negative comments about the decisions and work of other team members” and “My judgement and decision-making ability is excellent in emergency situations” carefully crafted to highlight both potential challenges and their solutions.
(The first question, for example, identifies whether employees are open to giving and receiving negative feedback from other employees, while the latter helps executives explore the prevalence of overconfidence in an organization.)
The dashboard is then capable of dividing results by department, and illustrating them in graph and table form in addition to delivering insights.
Kudashkina is confident the platform, which has been designed to be industry agnostic can effectively support any company with at least 15 employees.
Presenters illustrate Canada’s “competitive edge,” VC says
Among the industry experts watching the startups’ presentations was Angela Tran Kingyens, a partner with San Francisco venture capital (VC) firm Version One Ventures, who told ITBusiness.ca she loved watching the presenters’ many creative applications of AI and machine learning.
“I think it shows that Toronto has a competitive edge in thought leadership,” said Kinyens, who was born and raised in Toronto, but has been living in the San Francisco Bay area for the past seven years.
“There’s nothing better than coming home and being around Toronto’s incredibly contagious, ambitious tech ecosystem, and being a champion for it globally,” she said. “Every time I go back to San Francisco, I get to promote what’s happening in our ecosystem here, and NextAI adds to that.”
However, while noting that she was “really excited” to see the dreams and ambitions of NextAI’s first cohort come to life, Kinyens also emphasized that her years of experience have taught her that now is when the hard work begins.
“Events like this are important… but now it’s time for us to actually do the work of building real companies that have real impact, because ultimately we’re going to be measured on the success of what we foster and cultivate, and not what we celebrate in advance,” she said. “I’m looking forward to that.”