Amir Moravej’s body may have been in Tehran, but his mind was in Montreal. The engineer had spent a half decade in Canada, but an expiring work permit forced him to leave the country and return to his native Iran. Back home, Moravej scoured immigration forums and joined group chats where applicants shared advice and information about their cases. “It was impossible for me to read all of it,” he recalls. “So I wrote a bot to go and read all the forum posts, and find the ones that were most relevant to my own case.”
Moravej had arrived in Canada in 2009, for a master’s degree in engineering at Concordia University. After graduating, he founded a messaging app, and tried to obtain permanent residence via the federal government’s much-touted Startup Visa program. But the scheme doesn’t cover Quebec, so Moravej applied under the provincial program for locally-educated students. By this point, his work permit was running down, so he returned to Tehran to await a decision on his application. Then came the forums, chat groups and the bot.
A little over a year ago, Moravej landed in Montreal once more, this time as a permanent resident. His creation has become Botler, an immigration tool powered by artificial intelligence. On Wednesday, Moravej and his team launched an updated product, designed to help users through the process of putting together an immigration application. The first scheme Botler is being applied to: the Programme de l’expérience québécoise (PEQ), for foreign workers and students residing in Quebec. “Imagine you don’t have any information about the [program] but you want to apply,” instructs Moravej. “You can use this bot for the whole process, from the very beginning to the very end.”
Users start by answering questions about their qualifications and circumstances, which allows Botler to determine if they’re eligible for the program. Would-be applicants who meet the criteria then upload their documents, which the tool reviews. “If everything is fine, the bot will create an application package” that can be submitted to the immigration ministry, explains Moravej. Users who miss the mark get to see what gaps remain in their application, and what conditions they must meet to become eligible.
Botler’s machine learning engine uses the guidelines published by Quebec’s immigration department, and was trained on anonymized data from real cases. Lawyers at Montreal firm Campbell Cohen, which is partnering with the startup, conducted the product’s quality assurance testing. If the pilot project proves successful, the bot could be applied to other immigration programs.
Immigration lawyers and consultants often charge in the high thousands of dollars to help clients through the application process. Moravej says Botler will cost a tenth as much for do-it-yourself applicants. Campbell Cohen will also charge a reduced rate for users who want to retain human counsel.
The dream of automating parts of the typically laborious immigration process is not a new one. Government immigration agencies host their own eligibility self-assessment tools, notes Sharry Aiken, associate professor at the Queen’s University Faculty of Law. “The one thing you can’t really figure out is whether your [supporting] documents are adequate, but frankly I would have a hard time understanding how a computer would be able to give a definitive opinion on that,” she says.
Economic immigration is a “very criteria-driven program” says Aiken. In recent years, the Canadian immigration regime has increasingly aligned with labour market needs. “To the extent that the program does that, discretion and case law and interpretation become much less of a factor.” That combination of freely available information and predictable outcomes is what makes immigration law comprehensible for an AI—but it also reduces the need for such assistance in the first place.
For the “vast majority of candidates,” says Aiken, the intervention of a lawyer or bot is unlikely to do more than an accurate self-assessment using government-provided tools. She cautions against companies advertising increased odds of application success. “The people for whom for any kind of value-added support is useful with respect to economic immigration applications are [those] who don’t quite fit,” she says. For such applicants, whose qualifications don’t clearly meet the criteria, “a lawyer or a paralegal can make the difference between a ‘yes’ and a ‘no.’”
Botler can’t handle such complicated cases right now. But plenty of candidates retain legal help because they’re intimidated by the system, or for peace of mind. “I’m an engineer—I thought that I could understand everything, but I couldn’t,” says Moravej. “Even if you’re not one of those special cases, immigration is still a very difficult process.”
Assisting with an application from start to finish is an evolution for Botler, which in its first incarnation was a souped-up version of Moravej’s original forum-scanning bot. “If you [wrote] a very brief bio of your immigration case, the bot could go and find the most similar cases,” he explains. The largest of those forums was canadavisa.com, which is run by Campbell Cohen, now Moravej’s partner firm. Moravej met with founding partner David Cohen, who suggested they collaborate. Around the same time, the nascent company attended the city’s StartupFest, where it won an award. That’s when Moravej decided to focus on Botler full-time. “We decided to think of the immigration process more fundamentally, instead of just finding information from forums,” he recalls. “I went through the process myself [and] hit a roadblock, and my [co-founder] had the same problem. So we knew the problems.”
The PEQ is Botler’s pilot project, but Moravej says the technology can be adapted to other programs. “We are building something which is completely immigration program-agnostic,” he says. “The criteria changes for each program, but the way that the machine basically processes or reviews your file doesn’t change.” The logical next step is the federal permanent residence scheme immigrants apply to once they receive their provincial nomination; Botler is hoping to have that up and running by the spring. The technology could in theory be applied to any immigration program with publicly-available rules. If it works, the next Moravej may not have such a hard time staying in his new home.
- This Toronto startup uses A.I. to help lawyers predict trial outcomes
- Beagle uses machine learning to point out red flags in contracts
- Meet the artificial intelligence tech your company can use today
- How Canada’s immigration policy is failing high-tech startups
- Canada could be an A.I. powerhouse—if we can halt the field’s brain drain
- IBM Chief Innovation Officer Bernard Meyerson on the future of artificial intelligence
- Venture capital funding is rushing into machine learning startups
- Robots will mostly take over the jobs humans hate doing anyway
- Artificial intelligence just isn’t as smart as everyone thinks
- Why robots aren’t ready to manage human workforces (yet)