Earlier this year, Toronto-based Float made waves when secured nearly $100 million in financing (following a a $70 million Series B in 2025), with much of that capital earmarked for product enhancements and expansions.
Today, the Canadian firm unveiled Float Intelligence, an AI-powered automation layer embedded across its platform.
More than 7,000 Canadian businesses use the fintech platform to manage payments, optimize cash flow, and close their books.
The initial tool from the Float Intelligence kit is a transaction coding agent that automatically assigns general ledger codes and Canadian tax codes to corporate card transactions.
What once required hours of manual line-by-line coding now takes just minutes, according to company cofounder Rob Khazzam.
“We didn’t set out to build an AI product,” says the chief executive officer. “We set out to remove the everyday friction caused by Canadian businesses being handed infrastructure that was never designed for them.”
“We spent six years fixing that foundation,” Khazzam said. “Float Intelligence is what becomes possible once it’s right.”
The launch arrives as Canadian businesses face a “deepening efficiency squeeze,” with a recent Float report finding that while revenue grew last year, profits declined, cash reserves dropped, and debt remained flat.
Float Intelligence aims to give these companies an advantage by returning operational capacity so finance teams can deploy resources more strategically.
The transaction coding agent, purpose-built for Canadian context, handles the complexity of provincial tax structures, the unique ways Canadian businesses categorize spending, and the high accuracy standards finance teams demand, according to a statement from Float.
Float’s architecture combines a large language model with a proprietary per-business calibration layer trained on years of Canadian transaction data. The self-improving model learns each customer’s accounts, vendors, and tax assignment patterns, as well as from every user correction.
For privacy and security, Float runs all inference in its own secure AWS environment. No transaction data or coding decisions are shared between businesses, and each customer’s model is trained on their own data.
Accuracy and trust “are non-negotiable for finance teams,” notes Ruslan Nikolaev, cofounder and Head of Product at Float.
“Because Float already sits at the centre of how these businesses manage their spend, the model learns from how each customer actually works,” Nikolaev says. And, “If it isn’t confident, it stays out of the way.”
The transaction coding agent is the first of several planned Float Intelligence capabilities, according to Nikolaev.
The agent is available immediately to all Float Professional plan customers.
Leave a Reply