Spade, the data and AI platform for modern financial services, today announced Spade Predictions, a product suite for banks and fintechs built on a transaction foundation model Spade trained on payments behavior and the verified merchant behind every transaction.

Spade built its transaction foundation model to serve the whole industry. More than half a dozen of the world’s largest financial institutions and payment networks have trained foundation models on their own transaction data in the past eighteen months. Each required tens of millions of customers, a dedicated research team, and years of governed transaction history, and served only the institution that built it. With Predictions, Spade offers the same capability to institutions of every size, from the largest banks to regional institutions and fintechs.

Card networks still allow just 25 characters for a merchant name, a limit set decades before digital-first banking. Spade’s model has two core components. One learns what a business is, from Spade’s own verified record of it rather than from the payment message. The other learns what a person’s spending says about them, reading their history with the real business attached to every transaction. Combined, they can be leveraged to solve problems across an institution, from which product a customer is ready for, to which merchant they are likely to shop at next, all the way to detecting fraudulent behavior. Spade never receives personally identifiable information, and a model tuned for one institution runs on that institution’s own customer identifiers.

“A branch banker used to know their customers,” said Oban MacTavish, cofounder and CEO of Spade. “They knew who was about to buy a house and who had just had a child. Banks lost that as they moved to a digital first model, and the transaction record they were left with was built to settle payments, not serve customers. When a bank looks at a transaction now, it says almost nothing about the payment, so a bank with fifteen products has no idea which one to put in front of their customer. A model that knows exactly where each transaction went, and what that business actually is, can tell them which of the fifteen matters, and when.”

Fine-tuned for a US bank’s lending program, Spade’s foundation model outperformed a traditional propensity model built for the same task by 35%. The version using Spade’s merchant embeddings found roughly 10% more of the customers who went on to borrow than the same model without them. Spade’s foundation, model fine-tuned to detect fraud rings, replayed eight months of a card issuer’s transactions. Of all the fraudulent spend that cleared in that time, it would have flagged the merchants early enough for the issuer to block up to $3 million of it. In one case the warning came five months before the issuer’s first recorded block.

Predictions is the third product suite built on Spade’s Merchant Graph, a proprietary, verified record of 99.9% of US and Canadian businesses matched in real time at over 99% accuracy in under 40 milliseconds. It holds what each business actually is, what it sells, where it operates, which other businesses it resembles, how customers engage with it, and how every location rolls up to the brand and owner behind it. It powers all three suites:

  • Enrichment: the real, verified merchant, location and brand on every transaction, across every payment rail, including the long-tail, local, new and third-party-routed charges other approaches miss.

  • Data Operations: register the merchants and categories that matter and Spade returns the institution’s own identifier on every transaction that hits them, keeping it current as merchants rename, move and multiply merchant IDs (MIDs),with no merchant list of their own to maintain. Teams ask questions of their own data in natural language and get answers in seconds rather than as a multi-week analyst request.

  • Predictions: from which product a customer is ready for and which segment they belong to, to whether they are more likely to spend at one merchant than another, to which merchants are running fraud, all of them drawing on Spade’s merchant-level embeddings and models trained at the user level together. The suite covers the foundation model itself, models tuned to an institution’s own objective, and the software, infrastructure and embeddings for teams that would rather build their own.

Spade can be accessed via an API or MCP server as well as through the Spade Console. To learn more, please visit https://spade.com.

ABOUT SPADE

Spade is the data and AI platform for modern financial services. It identifies the real merchant on every transaction, across every rail, and pairs that data with its own verified record of 99.9% of U.S. and Canadian businesses and how people spend at them. With Spade, fintechs and banks can cut the cost of charges customers don’t recognize, win more of their customers’ spend, and turn their first-party data into new revenue and sharper decisions. Customers including Bilt, Mercury, and Stripe use Spade’s data to power rewards programs, spend controls, sharper analytics, and the models that turn transaction data into revenue.

Media gallery

About The Author
Business Wire