Markets Stocks Economy Crypto Earnings Banking Energy
Home› Tech› Feature
Tech · Exclusive

Propel's AI cash-flow checks could widen US credit access by 2027

Propel's AI cash-flow checks could widen US credit access by 2027
Tech · 2026
Photo · Eleanor Whitfield for Daily Digest Invest
By Eleanor Whitfield Markets Editor-in-Chief Nov 18, 2026 4 min read

Propel Holdings, a Canadian online lender, says it has developed a new way to judge creditworthiness that could open the door to loans for millions of people who are currently overlooked by traditional credit scores. The company's "cash-flow underwriting" model, built with data-sharing network Plaid, scans roughly six months of a borrower's bank transactions in seconds to assess income stability, spending habits, and existing financial obligations.

The system is set to roll out in the United States in early 2027, and Propel believes it could expand access to credit by as much as 15%. But the real test, analysts say, will be whether the approach can keep loan losses in check as it scales.

How cash-flow underwriting works

Traditional credit checks rely heavily on credit bureau scores, which are built from a person's history of borrowing and repaying debt. That system works well for people with long, clean credit histories, but it can leave out a large group of consumers: those with "thin" files—few or no credit accounts—or those whose past mistakes don't reflect their current financial reality.

Propel's model takes a different approach. With the borrower's permission, it pulls bank transaction data through Plaid, a service many people already use to link bank accounts to apps like budgeting tools or payment platforms. The AI then analyzes that data to spot patterns: regular paychecks, consistent spending, and whether the person is already stretched by existing bills.

The idea is that a person with a low credit score but a steady job and manageable expenses may actually be a safer bet than their score suggests. Conversely, someone with a decent score but erratic income and heavy spending might be riskier than they appear.

This kind of underwriting is not entirely new—some fintechs and lenders have experimented with bank data for years—but Propel's use of AI to process the data in seconds, and its plan to deploy it at scale in the US, marks a significant step.

Why it matters for borrowers and investors

For consumers, the potential benefit is clear: more people could qualify for credit, and at fairer prices. "Adverse selection"—the tendency for the riskiest borrowers to be the most eager to take a loan—has always been a challenge for lenders. If cash-flow data can identify borrowers who look weak on paper but are actually reliable, lenders can approve more loans without taking on outsized risk.

For investors in Propel, the math is straightforward. The company's claim of up to 15% more credit access only translates into profit if the new borrowers repay their loans. If delinquency and charge-off rates—the portion of loans written off as uncollectible—stay stable, Propel can grow its US business without needing to raise prices or set aside much larger loss reserves. That, in turn, affects how cheaply it can fund new lending.

Investors will likely focus less on the speed of the transaction scan and more on whether early US cohorts perform well enough to support growth at a reasonable cost. The company will need to show that its AI-driven model is not just fast, but accurate.

Broader context: AI in lending

Propel's move comes as the financial industry increasingly turns to artificial intelligence to make lending decisions. From Microsoft's cheaper AI models to Alibaba and Meituan using GPT-4 to train smaller models, AI is being applied across sectors. In credit, the promise is that AI can find patterns in data that humans—and traditional scoring models—might miss.

But the use of AI in lending also raises questions about fairness and transparency. Regulators are watching how algorithms make decisions, and lenders must ensure they are not inadvertently discriminating against protected groups. Propel will need to navigate those concerns as it expands.

The broader credit market is also in focus, with investors weighing opportunities in credit investments that offer both income and capital gains. And the growth of private credit has drawn warnings from regulators about potential blind spots. Propel's model, if successful, could become a template for other lenders looking to use alternative data.

What to watch next

Between now and the early 2027 US launch, investors will be watching several things. First, how Propel's model performs in its current markets—does it actually reduce losses? Second, whether the company can secure the funding needed to scale its US operations. Third, how regulators respond to AI-driven underwriting.

For everyday investors, the story is a reminder that the way credit is granted is changing. As AI and data-sharing become more common, the definition of a "good borrower" may shift. That could mean more opportunities for consumers—and new risks for lenders that get it wrong.

Propel's bet is that better data leads to better decisions. The next few years will show whether that bet pays off.

More from this story

Next article · Don't miss

Amer Sports sales surge 29% but stock lags; buyback signals value

Arc'teryx and Salomon owner Amer Sports is growing fast—sales up 29% this year—yet its stock is down 17%. With a $1 billion buyback and improving margins, the market may be missing the story.

Read the story →
Amer Sports sales surge 29% but stock lags; buyback signals value