Skip to content
Microsoft Industry Blogs

Learn more: request a free Microsoft Azure Government trial

In Part 2 of our series on how Microsoft partner Pondera Solutions is helping state governments combat fraud, waste and abuse, we learn why Azure Government was selected to power its advanced prediction algorithms. Pondera’s founder and CEO Jon Coss also shares some compelling examples of fraud-detection results since forming the company in 2011. In case you missed Part 1, please click here; otherwise, enjoy Part 2!

Can you elaborate on why Pondera selected Microsoft Azure Government to deliver the cloud power for your solutions?

If a technology is already available to do what we need, then we’re not going to build it … We apply our intellectual property more on the fraud algorithms and fraud-detection services … We were building our stack using Microsoft for a number of years and decided to build our constraint-based modeling system on Microsoft. We think this system is critical to detect fraud patterns, both existing and emerging, in a way that traditional predictions tools cannot do. But this requires us to perform huge numbers of calculations in a short period of time—and Azure can accommodate this.

Then you (add) the work you are doing in artificial intelligence, Microsoft Azure Machine Learning, Azure Data Lake, etc.—those are all things that we can apply to this problem … We literally have this seamless platform and our clients don’t have to worry about the underlying technologies. They know that it’s secure, they know that it’s scalable, they see evidence of it being integrated and that we’re passing these very high-quality, investigation-ready leads to them. I think that you’re quite possibly the only company in the world that could have accommodated us!

Jon, can you provide a few examples of outcomes achieved for your government clients?

To date, our system has identified well over $5 billion in likely fraud; our clients have collected or prevented hundreds of millions (of dollars); we’ve seen dozens of indictments … We don’t talk about a lot of our clients’ cases … but some clients talk about their cases specifically. For example:

  • One fraudster had set up a fictitious business (and) had actually stolen a number of IDs from people that were using student visas. The “business” was set up to hire employees using these fake or stolen IDs. Then, shortly after hiring them, they were fired and unemployment insurance claims started to get filed against the business from about five states … The one state that had our system was able to pinpoint what this person was doing and why they were suspicious … so they ended up prosecuting the case.
  • We also helped bust a scheme where a couple of folks in a Southeast state had created a business in the Midwest. They started doing tax claims against that business, obviously for refunds, and the refunds were being sent to homes that were either foreclosed or for sale because no one was there to pick up the mail. They were likely working with a real estate agent that was telling them “here’s where you should send them” and then literally following the mail truck around to pick up the checks.
  • We had one state that went to an industry trade show and said that, in the first three months of turning on our system, we identified $117 million worth of what they called high-collectable, high-yield fraud, waste and abuse.
  • We had another state that when we turned the system on, they got more than 300 examples of fictitious businesses that were operating in the state that the state then closed down.
  • Another program … using our system in combination with their detection team suspended about 30 percent of the Medicaid program providers, took 40 percent of the dollar flow out and issued a number of indictments.
  • Where we do case management and we can directly measure results, typically our return on investment is just around 300 percent on detection—and that does not consider prevention, which would measure how long would someone have been able to operate if our system wasn’t in place to stop them?
  • The opioids (crisis) is such a hot topic. Our system has helped states identify doctor shopping, pill-mill activities, those types of things. In some cases, we actually found patients that in a single quarter were receiving upwards of 200 prescription medications from different providers and pharmacies in a state … Most of those drugs that were obtained through the Medicaid program end up being sold on the streets.

How should prospective Pondera customers learn more?

They can certainly visit our website and I’d encourage them to watch our product videos and videos about some of our investigators. But we also like to put people in touch with one of our clients, because they’re our best salespeople … Then they learn more about techniques, about what has and hasn’t worked, how they use the analytics.

When they talk to us, they’re going to talk to somebody who came out of state programs. We hire more than half of our employees directly out of government; we have former FBI agents, state law-enforcement officers, data-mining specialists, Medicaid fraud control unit officers, unemployment insurance fraud investigators.

What about the future of fraud detection?

I think we’re heading in the right direction. This focus on investigation-ready leads should make companies like us a lot more accountable … Certainly this focus on sharing techniques and even data across states and programs. We’re all in this together, and if we build this club of vendors, of states, of programs and we look at this problem together, then what I’d really love to see is that we start talking about eliminating fraud instead of fighting fraud.

Anything else?

Microsoft seems to have made a real shift … The technology and Azure and the things you do to support us are great, but the fact that the company is behind us is really important … You’re trying to address real-world problems … that’s exciting to us … Fraud, waste and abuse in government is just a massive problem. We appreciate that Microsoft sees this and isn’t solely focused on selling us more Azure computing power. Instead, you’re helping us get the word out to more people about just how serious a problem this is. That kind of support from Microsoft is a really big deal to us.


My thanks to Jon for sharing his insights. We’re humbled to support such an innovative company that’s providing world-class services to governments—and benefiting taxpayers. Please: