Perspectives

Future of investigations

Four dimensions of the analytics-driven approach

By employing advanced analytics approaches in combination with field-demonstrated forensic techniques, organizations can better detect, isolate, and deter fraud attacks, with potentially significant positive impact on an organization’s performance and productivity.

May 08, 2018

A blog post by Satish Lalchand, principal, Deloitte Transactions and Business Analytics LLP.

Protecting data, intellectual property (IP), and finances have become an increasing priority at the boardroom level as fraudsters proliferate and constantly adapt to more sophisticated controls and monitoring. Even organizations with antifraud controls can have their investigative efforts impeded by several factors like reliance on rules-based testing, information silos, and volumes of unstructured data. But the traditional, rules-based approach to fraud analytics is shifting to be more integrated and cognitive. This new approach can help identify where to look for problems, unlike an intuition-driven approach that is purely based on the experience and knowledge of the investigators.

Potential benefits of the analytics-driven approach

Instead of writing dozens of rules or queries, investigators can focus on using their forensic investigative skills and experience to examine the narrowed down population, for instance, of items purchased in an effort to identify the few that warrant a closer look. This approach can save considerable time and more accurately hone in on potentially troublesome activities. It also can result in fewer mistakes while supporting more thorough analysis—a machine can help identify a trend that can often get overlooked by a tired pair of eyes.

Four dimensions of the approach

This integrated, analytics-driven approach encompasses several dimensions:

  1. Analytics maturity. The ability to conduct an analytics-driven investigation begins with determining where an organization resides on a maturity model that captures the people, process, and tool dimensions of fraud analytics and forensics. Factors contributing to an organization's analytics maturity include:

    The frequency with which the organization conducts analysis;
    Analytics tools being leveraged; and
    Whether analysis is conducted in silos or in an integrated, enterprise-wide manner.

    Functions such as marketing, customer experience management, and supply chain, which typically have strong analytics operations, could be sources of assistance and resources in ramping up analytics capabilities as part of an investigation.
  2. Integrated data marts. The ability to integrate structured and unstructured data from internal and external sources into risk models is fundamental to an advanced analytics response. Structured data alone provides a limited view of patterns that might point toward fraudulent activity. Likewise, when data is only available in organizational silos, the links between potential patterns may be hidden. An integrated approach brings together structured and unstructured data from across the enterprise, along with data from external sources such as watch lists and social media, to present a broader picture of activities and transactions, which experienced forensic investigators—aided by advanced analytics—can piece together with fewer false positives.
  3. Risk-scoring of the entity rather than the transaction. Transactions doesn't commit fraud. Employees, vendors, customers, and others do. Data-driven advanced analytics models incorporating text analytics and network analysis enable organizations to rank risks at the individual or entity level, rather than the transaction level. This approach, which incorporates statistical concepts rather than arbitrary risk ranking, can provide a broader picture of what is happening with an entity than analysis conducted on a test-by-test basis. Letting the data "talk" instead of subjectively assigning risk scores can improve ranking accuracy and efficiency.
  4. Application of predictive tools. Advanced analytics techniques, such as machine learning and cognitive computing, enable the study of transactions associated with bad actors. Insights into fraudster attributes gained through this analysis and reinforced by the knowledge and experience of forensic investigators can be used to "teach" models in an effort to identify individuals or entities exhibiting the same or similar traits in a broader population. Machine intelligence and computer decisions through artificial intelligence are starting to take precedence in detecting the digital footprint left behind by fraudsters. Development of this capability is a significant step in the maturation from reactive to proactive fraud analytics, helping to elevate compliance from a "man vs. machine" team to more of a "man and machine" team.

The continually growing appetites and capabilities of fraud perpetrators suggest that traditional models need to shift. By employing advanced analytics approaches in combination with field-demonstrated forensic techniques, organizations can better detect, isolate, and deter fraud attacks, with potentially significant positive impact on an organization's performance and productivity.

Stay tuned for our next post on the five ways to overcome data challenges in forensic investigations.

Insert Custom CSS fragment. Do not delete! This box/component contains code needed on this page. This message will not be visible when page is activated.
+++ DO NOT USE THIS FRAGMENT WITHOUT EXPLICIT APPROVAL FROM PLATFORMS TEAM +++

Site-within-site Navigation. Do not delete! This box/component contains JavaScript that is needed on this page. This message will not be visible when page is activated.

Did you find this useful?