Evolving Government: Battling fraud with big data

(Getty Images)


Written by

Part 4 of the Evolving Government series with Dcode42

Financial fraud takes many forms across the civilian agencies and costs these agencies collectively over $500 billion dollars every year. The issue crowds out potential government investment in other mission goals and is one that the federal government has tried to tackle government-wide with limited success. The intention is well directed, but the current approach makes it very expensive and slow for these agencies to detect, prevent, investigate, prosecute and recover the monies lost or misplaced.

There are agencies that are starting to take a new and innovative approach to this problem—one that has shown great promise in being able to significantly reduce the time and resources involved in the detection, prevention and investigation of fraudulent activities.

Financial fraud involves several important individual data points: accounts, individuals, businesses in all their various structures, those entrusted to facilitate the determination and payment of monies owed and the institutions receiving the funds.

These data points on their own convey an incomplete picture; one that leaves those tasked with discovering, rectifying and preventing fraudulent activity struggling to efficiently and promptly make sense of all the data. AtomRain was able make sense of all of these disparate data points at a lower cost and with less effort than current tools provide.

Using our product, GraphGrid, we connected all those data points together with contextually relevant relationships that make the picture complete; even more than that we enhance the picture because you see more than was there originally. We enable complex questions and problem statements to be asked and answered in real-time. An example problem statement could be the following: “starting with a given entity that is being transacted, show me the individuals and businesses that benefit the most as well as those most likely to have conflict of interest before and after the transaction based on the legal ownership structure above and below this entity; let me take this starting result set, visualize the network and explore the details from there to complete my analysis in real time.”

Below is an extremely simplified example of the structures that exist. Here we see an Acquiring Corp with two proposed acquisitions. One for a Corp owned by Sue and Sam and another option for one owned by Bob and Bill. The interesting thing that emerges is the potential conflict of interest in choosing which company to acquire where we see that Bob has managed to obtain a controlling ownership in Acquiring Corp and can now influence the acquisition to give him a payout on the sale of his other company Transacted 1 Corp. Returning and visualizing such a structure is a sub-second request and truly transforms the way an analyst works.

The issue illustrated above is not in the access to the data. The issue is that data as it is structured, stored and processed today cannot answer the questions that need to be asked and answered in real-time.

Financial fraud and data challenges aren’t relegated only to federal. Over the last 5 years AtomRain has been involved in introducing innovative data solutions to highly regulated, typically slow to change, global financial institutions from exchanges to banks.

GraphGrid is a transformational platform that delivers a production-hardened connected data architecture that supports extremely complex and highly-connected missions; providing out of the box services and frameworks for disparate data source unification, knowledge representation and learning, search and discovery, analytics and recommendations, distributed compute and human in-the-loop decision making and investigation workflow mission support.

The common thread we’ve found is that the challenge with these institutions has not been a lack of desire to be innovative and do more. The challenge is the tug-of-war between those tasked with defining, designing and implementing standard processes that keep an organization compliant and those that are given the mandate to innovate, grow and improve faster than competitors.

We like to restate the mandate a bit differently to unify it for everyone within the organization to tug in the same direction: We need to innovate in a rapid yet reliable, drastically different yet predictable and compliant, unknown yet measurable manner without putting our current obligations at risk.

This may sound like an impossible unification and when trying to execute it all internally it likely is; however, having an external team with deep expertise leading innovation initiatives from ideation to implementation and production operation to organization change strategy lead you through the process of introducing something drastically different while proving measurable results and value at each step makes succeeding with your innovation mandate possible.

The checks and balances, regulation, process, visibility, accountability and tug-of-war in a federal agency is just a different side of the same coin. Seeing the impact and innovation that has been possible for us to lead within the highly-regulated commercial financial sector, gives us the confidence and experience needed to help federal agencies navigate the introduction of innovative technology solutions that drastically change for the better the way they approach their mission.

The shape of data, big or small, within all federal agencies is a graph and it needs to be modeled, stored and computed as such to fully maximize its value as a mission asset. The beauty of the graph technology we provide is that you don’t need millions of data points to train a machine learning model before you can start getting value. Sometimes it is the “small data” (google it) that holds the key. If you do have big data with hundreds of billions of data points and constant streams flowing in that need analyzed, understood and acted upon in real-time, we’ve got you covered there too.

AtomRain went through Dcode42’s Big Data & Advanced Analytics accelerator program earlier this year, and because of the extensive exposure to federal partners and challenge areas, AtomRain has developed a comprehensive understanding of the data assets and utilization across the government.

See the current cohort of Dcode42 companies at the Artificial Intelligence and Machine Learning Government Demo Day on July 25 to understand how these types of emerging tech solutions can solve many mission critical agency problems.

Ben Nussbaum is co-founder of AtomRain. 

-In this Story-

DCode42, evolving government
TwitterFacebookLinkedInRedditGoogle Gmail