Salad Finance built its entire model around credit invisibles: thin-file consumers, thin-file merchants, and people PSPs routinely decline at onboarding. Most operators read that as a niche fintech play. I read it as a structural indictment of how PSPs and EMIs underwrite.

As Frederic Yves Michel NOEL highlights, the standard onboarding stack still leans on credit bureau data, bank statements, and historical processing volume. If a merchant lacks two of those three, the risk team says no. That is not risk management. That is a refusal to build a real underwriting model. Meanwhile, thin-file merchants are not unprofitable. They are simply unmeasured. A PSP that cannot price or underwrite them is leaving margin on the table and pushing those merchants toward whoever will take them first.

For PSPs, the key point is this: the cost of saying no is not zero. It is the lifetime processing volume of every merchant you declined because your onboarding flow could not handle a thin file. Salad Finance did not invent a charity case. They found a segment the incumbents were too rigid to serve.

For PSPs and EMIs onboarding thin-file merchants today, where does your underwriting actually break first: identity, fraud scoring, or expected volume?