In the winter of 2022, as the tech world was reeling from the sudden, explosive arrival of Openai’s chatbot, Benjamin Alari An important choice was faced. His legal tech startup, Blue Jwas a venerable business built on AI from a bygone era, serving hundreds of accounting firms with predictive models. But it had hit the ceiling.
Alari, a Professor of Tax Law But University of Torontosaw the nascent, error-prone, yet powerful capabilities of large language models not as a curiosity, but as the future. He took on a high stakes: to uproot his entire company, which had been painstakingly built for nearly a decade, and rebuild this unproven technology from the ground up.
That bet has paid off handsomely. Blue Jay has since kept quiet 2 122 million series d Funding round by Oak HC/ft And Neelam Venturesmaintaining the company’s value More than 300 million. The move transformed Blue Jay from a niche player to one of Canada’s fastest-growing legal tech firms, growing its revenue nearly twelvefold and attracting 10 to 15 new customers each day.
The company now serves more than 3,500 organizations, including the global accounting giant KPMG and several Fortune 500 companies. It’s tackling a key obstacle in the professional services industry: a severe and perishable talent shortage. America has 340,000 fewer accountants than five years agoand with 75% of current CPAs expected to retire in the next decade, firms are desperate for tools that can increase the productivity of their remaining professionals.
“What once took tax professionals 15 hours of manual research to do, can now be done in about 15 seconds with BlueJ,” said Ellery, the company’s CEO, in an exclusive interview with VentureBeat. "That value proposition – we can take hours of work and turn it into seconds of work – is what’s driving a lot of it."
When Dean’s Biography Was Wrong: The Moment That Changed Everything
Alari vividly remembers January 2023, when the dean of the law school stopped by his office to say New Year’s greetings. He asks her about chatput and prompts the AI to explain it. Chet GPT developed a biography with confidence. Some details were correct. Others were completely fabricated.
"She was like, ‘Well, it’s really kind of scary. It’s wrong, and it has implications,’" Alari said. Yet this moment of apparent failure did not deter him. Instead, it crystallized his punishment.
The company’s first iteration, launched in 2015, used supervised machine learning to develop predictive models that could predict court outcomes on specific tax cases. While technically sophisticated, it had a fundamental flaw: it could not answer every tax research question.
"The challenge was that it could not answer every tax research question, which was really the Holy Grail." Alari said. Users loved the tool when it applied to their problem, but would quickly abandon it when it didn’t. Revenue was about $2 million dollars a year.
Despite Chetgupt’s notorious deception, Ellery convinced his board to pivot. "I was convinced that if we continued on this path we would not be able to deal with our number one limit," He said. "Large language models seemed like a very promising direction."
He gave his team six months to deliver a working product.
3 million questions from a 90-second response: How Blue Jay understood AI deception
By August 2023, Blue J was ready to launch. What they did issue was, in Alari’s clear assessment, "Super Junkie" The system took 90 seconds to respond. Half of the answers had problems. Net Promoter Score Registered in 20 only.
What turned that poor product into the platform it is today—with response times in seconds, a dissatisfaction rate of just one in 700 inquiries, and an NPS score in the mid-80s—was an intense focus on three strategic pillars.
The first is largely proprietary content. Blue J With exclusive licensing reserved Tax Analyst (Tax Notes) And IBFDthe Amsterdam-based global tax authority covering 220+ jurisdictions. "We are the only platform on Earth that has the best US tax information from Tax Notes and the best global tax information from IBFD." Alari said.
The second is deep human skills. Blue Jay is headed by tax experts Susan Masseywho spent 13 years IRS Office of Chief Counsel As Branch Chief for Corporate Tax. His team constantly tests the AI and improves its performance.
The third is an unmatched feedback flywheel. With over 3 million tax research queries processed in 2025, BlueJ is collecting unprecedented data. Each query generates feedback that is fed back into the system.
Weekly active user rates hover between 75% and 85%, compared to 15% to 25% for traditional platforms. "A welfare ratio is such that we use five times as much intensity," Alari noted.
As part of Blue Jay’s early access partnership with OpenAI
Blue Jay retains one An unusually close relationship with Openei This has proved crucial to its success. "We have a great relationship with Openei, and we get quick access to their models,"Alari said. "It is quite collaborative. We give them really high-quality feedback on how different versions of upcoming models are better."
This feedback proves valuable because Blue Jay builds on what he says "Environmentally sound" Test Questions – Created from original tax professional questions, with correct answers by Blue Jay’s expert team. This helps OpenAI improve performance on complex reasoning tasks.
The company tests models from all major suppliers. Open Eyefor , for , for , . Anthropicfor , for , for , . Google’s Geminiand open-source alternatives—continually evaluating which ones perform best. "We are not necessarily 100% committed to a particular provider," He explained. "We are checking all the time."
This approach helps Blue J Navigate a challenging business model: charging about $1,500 per set annually for unlimited queries while absorbing variable compute costs. "We offer them a really good user experience, unlimited tax research answers at a fixed price," Alari said. "We are absorbing a lot of that risk."
The foundation model shows that competition between providers creates downward pressure on API prices, while Blue Jay’s conservative usage modeling has been proven to be accurate. Gross revenue retention exceeds 99%, while net revenue retention reaches 130%—considered the best standard for a Sass business.
Competing Thomson Reuters and Lexnexus with 75% weekly engagement
Blue J As the competition competes with established publishers Thomson Reutersfor , for , for , . lexisnexisand Bloombergthey all announced AI capabilities in 2023 and 2024. Yet Blue Jay’s engagement metrics show it has captured significant momentum, growing from just 200 users to more than 3,500 organizations in 2021.
Daily updates prove to be important. Although the tax code itself changes as Congress acts, the ecosystem is constantly evolving through IRS regulations, new rulings, and court cases. All 50 states change their tax code regularly.
"Things are literally changing every day," Alari said. "Every day we are updating the content, and it is only the US that we cover Canada, we cover the UK. The aspirations for this thing are truly global."
Alari’s ambitions go beyond building a successful startup. As the author of an award-winning book "Legal uniformity" and affiliated with the faculty Vector Institute for Artificial Intelligencehe has spent years considering the long-term impact of AI on the law.
In academic papers published in Tax Notes 2023 And 2024he chronicled the rise of generative AI, predicting it "Clients will become quite sophisticated" And that AI will push human experts into high-value strategic roles rather than routine research.
Blue Jay’s 2 122 Million Project: From Tax Research to ‘Global Tax Awareness’
Series D Fundingwhich brought total capital to more than $133 million, will drive aggressive geographic and product expansion. Blue Jay already operates in the US, Canada and the UK, with plans to eventually cover 220+ jurisdictions through its IBFD partnership.
Future capabilities may include automated memo generation, tax form completion, document drafting, and conversation history that preserves context across sessions. "Operating layer for global tax realization."
For all its success, Blue Jay operates in a domain where mistakes have serious consequences. The problem of deception has not been eliminated – it has been minimized through careful engineering, material support and human supervision. Blue Jay trains its models to recognize when they can’t answer a question rather than fabricate information.
Businesses also face financial risks if calculation costs spiral or usage patterns exceed estimates. And subtle questions about professional judgment: As AI systems become more capable, will users defer output without significant critical evaluation?
15 Hours to 15 Seconds: What Blue Jay’s AI Axis Teaches Every Industry
Blue Jay’s transformation offers lessons beyond tax software. A company’s willingness to abandon eight years of proprietary technology and rebuild on an initially unreliable foundation required both courage and calculated risk-taking.
This decision paid off not because generative AI was inherently superior to supervised machine learning in all dimensions, but because it solved the right problem: comprehensiveness rather than accuracy in narrow domains. Tax professionals did not need 95% accuracy on 5% of questions. They required 100% accuracy on the questions well.
The improvement of NPs from 20 to 84 in just two years reflects relentless iteration informed by massive data collection. Content partnerships have created differentiation that pure technology cannot replicate. A team of tax experts provided the necessary domain knowledge to ensure reliability.
Essentially, Blue Jay recognized that the real competition wasn’t other AI startups or even established publishers. It was the old way of doing things – 15 hours of manual research, institutional knowledge in retired heads of professionals.
"People are like, ‘What does Blue Jay do? They provide better tax answers. Well, I think we need that,’" Alari reflects.
As AI transforms profession after profession, clarity of purpose may matter more than technical sophistication. The future does not belong to those who develop the latest AI, but to those who effectively use it to solve problems that humans actually encounter.
For a tax law professor who started out frustrated about inappropriate research methods, building a $300 million company marks a brave end point. Thousands of professionals are now answering complex questions in 15 seconds instead of 15 hours, representing the future of their profession, arriving faster than expected.
Betting on Chat GPT while it was still fueling biography has become a confirmation that sometimes the riskiest move is not to move at all.