Funding is the need a growing brand can't put off. It connects us to how the business runs, and shows us how the invoices we fund actually end. We are building our AI CFO on that record.
Last edited 5 min read
The first product a business buys from you decides what you learn about it. Sell a dashboard and you see what the customer chooses to show you. Fund the customer and you see what the decision needs, and then you see how it turned out.
Capital first is our go-to-market. Funding gets Spring in the door, and what funding shows us is what we build software on.
The need that cannot wait
A consumer brand that wins a retail order pays for the goods long before it is paid for them. Its manufacturer wants paying up front or on short terms. The retailer pays 60 to 90 days after delivery, and in food and beverage those terms have stretched toward 120.1 Add the weeks the goods spend being made and shipped, and a growing brand can have its cash tied up for three months or more in an order it has already won.
A founder can put off new finance software for a year. They cannot put off paying for the next purchase order. Funding is the need a growing brand can't defer, so it is where we start.
Funding earns the connection
Any app a brand connects can read its bank feed, its books and its storefront. To fund a brand well we go further: the retailer and distributor portals it sells through, and the EDI feed its purchase orders and invoices arrive on. Nearly every business we fund has its bank account connected.
Those connections stay live, because the next funding request runs on them. Software sold first has to earn each connection on its own. Funding arrives with a reason to connect.
Every invoice we fund ends in an outcome
When we fund an invoice, the retailer is usually told to pay Spring instead of the brand. That is the notice of assignment, and it means the payment comes to us. So for the invoices we fund, we see the result directly: whether the retailer paid, on which day, how much it held back, and the reason it gave.
The books record the invoice and, once someone matches it, the deposit. The reason a payment came in short usually sits in the retailer's portal, one line at a time.
Those short lines are deductions, taken for promotions, shortages and late deliveries. Estimates vary, but they can take a few cents of every dollar a brand sells.2 A brand can't plan its cash without knowing how much of that will come back.
Many of the brands we fund sell to the same retailers, so we see, in aggregate, how a retailer actually pays: how long it takes and what it tends to deduct. That goes into our next funding decision. No brand sees another brand's data. What pools is the retailer's behavior.
Why this order
Square and Shopify found this link from the software side: they began making funding offers to their sellers from the sales data their platforms already held.3 We run it the other way round, because sales and bank data are the easy part. What is hard to come by is the outcome: whether each invoice was paid, when, and how short. A deduction tool reads the short lines in the portal, the books record the deposit, a forecasting tool watches the bank balance, and someone joins the three by hand, if at all. A funder has to join them, because the money is owed to it: we match each payment to the invoice it settles and each short line to its reason, or we don't know what we are still owed.
What we are building: an AI CFO
A founder has questions no report answers. Will the cash hold through the next order? Which retailer is paying more slowly this quarter? Which deductions are worth disputing? How much more can we take on, and when?
We already forecast cash for some of the brands we fund, and we check those forecasts against the cash that actually arrived. Next we are building an AI CFO: pure software, always on, reading the same live data our funding runs on, and telling an owner what changed, what is likely to happen and what to do about it.
Connectivity comes out of our stack first. It already reads the bank accounts, books, storefronts, retailer portals and EDI feeds of the businesses we fund, along with how the invoices we fund actually get paid.
A new layer, built on that data instead of on funding. It turns what the data says into insights, likely outcomes and suggested actions. A brand will be able to use it whether or not it takes funding.
We are building four: cash flow forecasts from live data and real payment timing; insights on what changed, like a retailer paying more slowly than it used to; deductions, each matched to the line and the reason it came off, so a brand knows which to dispute; and capital planning that shows what the business can take on, and when.
The stack re-forms on a wider base. Funding and the AI CFO share the rails, the ledger and the data. A brand we fund will get the AI CFO on connections it already has, and every invoice we fund gives it another real outcome to check its forecasts against.
Funding is how we start with a brand, because it is the need that can't wait. What we learn while funding it goes into the software we build next.
1 Kristen Palmer, First Business Bank, writing in New Hope Network, September 2026: "Sixty- to 90-day payment terms are standard in retail," and in food and beverage they have stretched toward 120.
2 There is no public industry-wide benchmark for deductions, and published estimates vary widely.
3 Square Capital launched in 2014 and Shopify Capital in 2016. Shopify bases eligibility on a merchant's sales and activity on Shopify.