[Sean Cash]

I agree with Isaac. There are so many fraud files that a computer can not detect some of these deals, in my experience. It sometimes can be detected by a merchant interview and an underwriters experience not a computer

Some companies wrongly classify the merchant interview as as "funding call". I've never been on an interview where I already got the job and just had to show up to confirm my name and to say that I agree to the work terms. Why would a merchant interview be any different?

The interview is an opportunity to grill the merchant albeit in a nice way and decide afterwards if everything lines up with the docs and background check. I'm all for algorithms and I think that's the future but once there's a standalone mathematical formula for approvals, merchants will learn to game the system. Wonga, the UK payday lender has suffered through 40% bad debt because people started applying for loans in other people's names and all sorts of things. The system couldn't tell the difference.

If you think merchants are brazen now with fraud (and they are) just wait until they learn that humans aren't even underwriting anymore. It will become a game of beat the computer.

[JayBallentine]

If you think merchants are brazen now with fraud (and they are) just wait until they learn that humans aren't even underwriting anymore. It will become a game of beat the computer.

There are API's out there that will pull bank records. The merchant enters his deets; "bank log in" and "bank password."

The bank and API's verify your token and your server receives data to parse, analyze, underwrite.

Mint.com used these API's before they were bought for mega millions, then they switched over to another provider who would prefer these API's not exist at all.

We are talking about bank grade security - merchants stand no chance.