TISC is fine-tuning AI for split-second decisions, starting on the basketball court

A split-second decision can be the difference between a win or a loss; a championship ring or a participation prize. Accumulated over time, these can make a franchise or a failure.

“The game of basketball is very noisy… Think of us as noise cancellation for decision making.”

Those are the stakes the Toronto- and San Francisco-based startup The Intelligent Search Company (TISC) is betting that sports teams will pay for. The startup is fine-tuning existing artificial intelligence (AI) models to instantly surface context and recommend actions—particularly for coaching teams.

The company, which is incorporated in Delaware (more on that later), has exited stealth and secured $2.1 million USD ($2.9 million CAD) in pre-seed funding, the co-founders told BetaKit on Tuesday. Silicon Valley’s OVO Fund (no relation to Drake) and Toronto’s n49p led the round, with participation from Montréal’s Panache Ventures.

TISC was founded in May 2024 by University of Waterloo graduate Mahbod “Moe” Sabbaghi and Arpan Bhattacharya. The two met as software engineers at the e-commerce company Wish.

“We’re both very neurotic about search and software experiences,” Sabbaghi said. “Finding information has always been a problem for us.”

It all sounds a little abstract, but it comes together on the basketball court, the co-founders say.

TISC is building its software to be the analyst that never sleeps, and that can filter through all data on a team’s performance: video footage, player stats, and league historicals.

“The game of basketball is very noisy,” Sabbaghi said. “There’s sources of information coming from every angle. Think of us as noise cancellation for decision making.”

The co-founders claim that current, general-purpose large-language models (LLMs) aren’t purpose-built for picking out the right actions from a pool of data changing in real time. The co-founders claimed they have trained niche, open-source models to be quick pattern recognizers in fluid situations.

For basketball coaches, this could mean the AI prompting a play suggestion in a pivotal moment of a game, by relying on historical knowledge and data from similar situations.

Alex Norman, managing partner at n49p, said in a press release that TISC has built “technology that will enable any human or AI to find exactly what it’s looking for from any source or data stream, no matter how fragmented or complex that data is or how abstract the search query is.”

So far, it has piloted its capabilities with NCAA Division I teams from the Virginia Military Institute and the University of North Texas. It’s also set to do its first larger-scale pilot with the Women’s Premier Basketball Association in the Bay Area.

TISC plans to hire machine learning talent, expand early pilots, and “build relationships” in target markets. It’s currently just three people: Bhattacharya, Sabbaghi, a machine learning researcher helping out, and an incoming intern from McGill University.

Sports is just the first “beachhead” for TISC, the co-founders say, and relatively low-stakes compared to the markets it wants to eventually target, like emergency response lines, wildfire fighting technology, and even defence contexts.

“Even if it’s only a small percentage of revenue down the line, we see it as an effective laboratory,” Bhattacharya said.

Despite backing from United States (US) and Canadian VCs, TISC is incorporated in Delaware, which Bhattacharya said is “standard” for companies like theirs planning to operate in both countries.

“The reason why we have one foot in Canada and one foot in America is because this enables us to put feet into both pools of talent,” Sabbaghi said.

They’re not alone; Colare CEO Nain Abdi explained on The BetaKit Podcast earlier this week that VC interest and opportunities brought his Canadian startup to Silicon Valley. This summer, Canadian tech leaders, including Cohere CEO Aidan Gomez, encouraged Canadian founders to reject US VCs’ pressure to incorporate in the US.

CORRECTION (11/26/2025): A previous version of this article incorrectly stated that TISC had two and a half employees. That was a joking reference by one of the founders; the article has now been updated to state it has three team members.

Feature image courtesy TISC.