It seems to me like this company's base value proposition, that is, for any company's monetary transactions with customers, give that transaction a risk assessment score based on whatever data about the customer is available.
That, in its simplest, most elemenatary, most basic form, is a good value proposition.
Companies would find value in software which could deliver that proposition (a related idea that comes to mind is if PayPal decoupled their fraud prevention software, and sold that software to businesses separately, then that software would deliver an equivalent value).
So the base value proposition of this company (assuming their software actually works) -- is a valid one.
What seems to have done this company in, however, (if the article is to be believed) is accounting fraud, which may have been preceeded by lack of customers, lack of revenue or expected revenue.
It's always an interesting question to me (correlation vs. causation-wise) if lack of sales triggers accounting fraud, or if accounting fraud just sort of happens on its own...
You know, here's an idea for the FBI or other related investigative agencies -- you could, based on previous cases, create a list of "risk factors", each of which would count towards an overall "risk score" -- of something being seriously wrong at a company.
Such factors would include, but not be limited to: How much debt a company has to investors, how much revenue is it making relative to that debt, what is its growth rate, what is the age of the CEO, what industry is it in, etc., etc.
Grab all of that data, from all companies, run it through a machine learning algorithm, get a "risk score" for each and every company, then if you have nothing better to do (free time between other investigations), start investigating the companies with the highest "risk scores".
You know, you might call it "Cyber Fraud Detection Software" -- for large companies...
Then if that works well in the U.S. -- give the software (for free!) to all other countries!
Also, the same set of ideas and Machine Learning -- could be used to track wasteful government spending (foreign and domestic), or wasteful spending of so-called charitable organizations...
The possibilities are truly endless!