Across all verticals, the repercussions of fraud are massive. In 2012, an interest rate-fixing scandal at a leading financial services firm spooked financial markets, tarnished the firm’s reputation, and garnered several hundred million dollars in fines. In high-tech and manufacturing organizations, obsolete corporate monitoring systems increase the potential for valuable intellectual property to be smuggled from patent-holding organizations to competitors. With internal threats more pervasive and costly than ever before, competitive businesses are increasingly prioritizing sophisticated fraud detection solutions.
To date, no big data platform has been equipped to efficiently manage the entirety of a company’s internal data deluge. The pure size of the data, speed at which it needs to be analyzed, and diversity of data types leaves potential red flags obscured within the fog. Some companies have attempted to address the problem with a relational database, which can handle merely a fraction of a company’s data while still becoming prohibitively expensive. Others have looked to Hadoop, an exceptionally insightful—but slow and complicated—platform. Firms on the “cutting edge” are leveraging RDBMS-Hadoop “connectors” to ship data back-and-forth between the two data stores; they are unnecessarily burning tremendous amounts of time and capital.
Hadapt fuses a relational database with Hadoop inside a single, connector-free analytics platform, enabling customers to work interactively with their multi-structured data via industry standard SQL. Consolidating data into a single, unified platform drastically reduces TCO, eliminates data silos, and allows for richer analytics.
The Use Case
A financial services firm leveraging multi-structured data conducts investigative analytics to detect and preemptively eliminate fraud. The company ingests terabytes of data per day from disparate systems (email, instant messages, audio, machine-generated events, etc.) to enable interactive analysis by business analysts. These capabilities also pattern-match for abnormal behaviors, signaling early warnings for substantial internal threats.
An additional example relates to insider trading, where customers want to examine the frequency and content of communication between employees and external agents. If communication spikes abnormally, for instance, an interactive console triggers an alert for a more detailed inspection. Analysts seamlessly sift through enormous volumes of multi-structured communications with SQL searches using full text search, natively embedded in Hadapt, to identify cryptonyms, suspicious phrases, and discussion of confidential information. Hadapt also natively integrates Mahout, an MPP machine learning system, within which customers have created SQL-based sentiment analysis algorithms via the Hadapt Development Kit™ to evaluate stress levels of phone calls. Analysts can therefore drill into communication rates, content, and sentiment—interactively, in one place, using standard BI tools.
Where there’s smoke, there’s fire. Fortunately, Hadapt has created the most advanced smoke detector in the world.