Discrepancy AI was named in the Top 28 Hottest Toronto Startups to Watch in 2025

Did you hear? Discrepancy.ai was recently named one of Toronto’s hottest startups by The Founders Press! The Founders Press is an online media platform dedicated to covering early-stage startup stories, bringing them to life, and sharing it with other entrepreneurs. The list is a spotlight on Toronto Tech Week, a citywide collection of events featuring the city’s fast-growing startup ecosystem. Discrepancy appears in spot #14!

Toronto is one of North America’s leading tech hubs, home to a diverse pool of talent and world class universities. With companies like Google, Amazon, and Slack located in town, Toronto offers great opportunities to entrepreneurs across the globe. 

We are certainly in good company among Toronto startups – The Founders Press highlights brands like Crafty Ramen, Athena AI, and sports tech pros at NTangible. We’re thrilled to be sandwiched between the recruitment experts at Linkus Group and fractional hiring whizzes at BrightIron! Participating in accelerators and community events has allowed us to learn from other founders and build a product that resonates with our customers and with other innovators. We couldn’t be happier to “share the spotlight” with such awesome startup teams.

Discrepancy was accepted into Techstars DC in fall 2024, and we continue to be a big supporter of the Toronto startup community and Toronto’s fintech ecosystem. More than 289,000 tech workers call Toronto home, and Toronto is the third largest tech hub in North America. There’s plenty of competition across industries like  fintech, healthtech, and insurtech – so we’re glad to have our mission resonate and stand out. 

Discrepancy’s Role in the Toronto’s Startup Scene

At Discrepancy, our focus is on making data easier: to pull, to read, and to use. Pulling financial data out of documents is difficult with legacy optical character recognition (OCR) systems, which use probabilities to predict labels and relationships. Newer methods, like Natural Language Processing (NLP) are better at seeing data in context, and require less manual oversight. The problem is that many companies have legacy OCR systems, and adding NLP on top just doesn’t fit. Discrepancy was built from scratch to solve these issues, process language in context, and serve up data that is fast and user friendly.

Identity documents for rental screenings, mortgage applications, or loans can look different in every single use case – in OCR, every type of bank statement, for example, requires its own model. With our current features, companies can save 80% on labour time, freeing up their expert team for higher value tasks. Rather than one single model, Discrepancy is built on models chained together; one main model distributes the tasks and the workload to smaller models. The result is a faster, more flexible system capable of harder tasks and more refined results. 

Thank you to The Founders Press for spotlighting our work, our peers on the list, and the Toronto startup world. We’re happy to live in a city where big ideas can thrive, and we can’t wait to see what comes next!