Prevent fraud and activity that may threaten your company’s bottom line. Discover hidden or latent threats with robust text analytics. Comb eData to meet regulatory and legal requirements. Gauge customer sentiment expressed in formal surveys, telephone calls, or unsolicited emails. The applications of CONQ span functions and activities that are critical to business success.
View CONQ applications for:
Criminal fraud has become an enormous challenge for both the commercial and government sectors; insurance companies face staggering losses. Property and casualty insurance fraud, alone, cost insurers tens of billions of dollars every year. Bank and credit card fraud are signature activities of gangs and terror-related groups.
Ordinarily, the insurance industry relies on frontline staff to spot fraud, which is often impractical given their workloads. Claims adjusters are expected to note suspicious indicators during five separate stages of the claim life cycle. Banks depend on tellers and managers, as well as very basic computer analytics, to discern unusual transactions. Banks and other industries are not equipped to detect fraud when the information is hidden in narrative data such as adjuster or teller notes, or suggested by subtle behavior patterns.
CONQ provides a solution through its patented ability to harmonize and fuse unstructured data and then apply that data along with structured data to analysis models called “Concept Systems.” These Concept Systems, which are combinations of ideas and terms that relate to the criminal activity, uncover otherwise hidden trends, patterns, and relationships.
By empowering the user through true behavior and text analytics, CONQ helps identify fraud with unprecedented speed and accuracy. The result is timely intervention, which can be critical to minimizing the harmful impact of fraud.
FRAUD DETECTION CASE STUDY
FRAUD DETECTION CASE STUDY
The Arlington, Texas Police Department needed to meet national accreditation requirements in a more timely, efficient, and effective way. It found the answer in CONQ’s text analytics.
One in ten insurance claims are fraudulent, costing the insurance industry tens of billions of dollars each year. At stake for the hardest hit insurance companies is not only their bottom line growth but their brand credibility. Such steep losses can get passed along in the form of substantial premium hikes, making their customers angry and their business less competitive.
Fireman’s Fund Insurance Company (FFIC), a property and casualty insurer based in Northern California, has an especially strong commitment to fighting fraud. One of the nation’s top 20 carriers, FFIC is protecting its strong reputation by becoming a leading innovator in technologies to detect, investigate and counter fraudulent activity.
The claims handling process in the insurance business is time consuming and labor intensive, typically involving five steps—investigation, assessment, evaluation, negotiation and settlement. Since suspicious activity can crop up in any of these stages, the adjuster handling the claim has to be on constant alert.
While claims adjusters have the training and expertise to spot fraud, they are overwhelmed by the rapidly growing caseload. Moreover, the evidence of fraud is often buried in vast amounts of text, making it easy to overlook critical signs of crime. When the adjusters do detect fraud, it is often too late for the company’s Special Investigative Unit (SIU) to follow up effectively since corroborating witnesses can be difficult to find.
To assist the adjusters, as well as the SIU team, Fireman’s Fund needed a technology that could quickly identify fraud in large volumes of unstructured text.
Knowing that vital information is embedded in adjuster claims notes, Fireman’s Fund turned to CONQ for its groundbreaking ability to spot hidden trends and patterns in textual data. FFIC analysts used CONQ to extract key “suspicion” concepts which they then combined with coded data such as claim date, location and amount. That resulted in patterns that typically correlate to fraudulent claims.
New claims were then evaluated against these patterns and scored for their fraud potential. If the system detected a strong presence of suspicion concepts, the adjusters were alerted.
Fireman’s Fund can now benefit in numerous ways from CONQ’s text analytics. FFIC analysts can identify claims likely to result in fraud with unprecedented speed and accuracy. The SIU teams can be more productive by spending less time reading through claim details and more time analyzing relevant data. And the company can address fraudulent claims in a far more timely fashion.
Fireman’s Fund is actively exploring using CONQ’s text analytics to better understand the “Voice of the Customer.” By analyzing customer feedback and open-ended survey responses, FFIC expects to gain insights into customer sentiments and identify ways to improve the customer experience.
“This tool is an innovative approach to a clear and present concern for all insurance companies.”
— Greg Castleman, SVP, Claims Field Operations
Risk analysis can be extremely difficult today, as analysts need to quickly identify critical information that is obscured in structured and unstructured data. CONQ is the groundbreaking solution. With our patented technology, CONQ enables users to extract the most pertinent information, whether embedded in databases, reports, documents, social media, emails, phone logs, or other data platforms.
When faced with suspicious activity, you need accurate and thorough information, and fast. CONQ delivers in several ways, beginning with software that harmonizes and fuses data while dramatically reducing complex leverage and time-consuming ETL processes. CONQ also allows analysts to their subject matter expertise by creating “Concepts”—ideas, terms, phrases related to suspicious or criminal behavior—that rapidly lead to discovery of relevant information otherwise hidden in vast amounts of data.
CONQ further enables analysts to combine these Concepts into sophisticated analytic models called concept banks, allowing expertise to be shared without the sharing of sensitive underlying data. By building upon each other’s knowledge, the analysts can far more effectively anticipate, identify and defuse future risks.
Businesses face a growing burden of complex laws and regulations that can require them to comb through eData that are largely in the form of unstructured text. Some situations that can call for eDiscovery are lawsuits, internal and external security breaches, and investigations of employees suspected of violating company policies.
According to a recent survey of IT and legal professionals, eDiscovery’s use and importance will grow exponentially in coming years, especially as the regulatory environment for companies becomes more and more complex.
CONQ provides companies with groundbreaking capabilities that make eDiscovery faster and far more thorough. By not requiring processes like data normalization and ETL (Extract, Transform, Load), CONQ reduces both the cost of eDiscovery and its burden on IT. Also, your company can use CONQ in the future as it establishes procedures to retain and access eData to comply with regulatory or legal discovery requirements. The expertise remains in and is shared through concept banks, without having to share the sensitive underlying data.
eDISCOVERY CASE STUDY
e-discovery CASE STUDY
Health Insurance Client
Facing allegations of unfair practice, a large health insurance company relies on CONQ to quickly, thoroughly, and effectively analyze massive amounts of eData to mount a defense.
A large health insurance organization had to quickly prepare its defense against charges of unfair practice. Consequently, this required the company to analyze 1.2 million call records gathered over a month’s time for any evidence of legal violations. These records included not only structured data but also unstructured narrative comments, making the task nearly impossible for manual readers.
The CONQ Solution
Within two weeks of installing CONQ’s patented software, the company managed to narrow down more than a million records to just 12 records of interest. This allowed analysts to review the distilled records manually, with speed and efficiency, in order to clear the company of the proposed allegations.
What made this possible were several revolutionary capabilities. First, CONQ’s patented learning process uncovered ideas, terms, and phrases within the data that related to unfair practice. Analysts were then able to create and refine “concepts” that could identify truly relevant information within the eData. Iterative analysis led to even greater insights, such as relationships, patterns, and anomalies that further narrowed the number of pertinent records.
The result: 1.2 million calls were exhaustively checked for evidence of unfair practice and, compared to previous approaches, it took fewer people far less time to achieve far more accurate results.
The health care organization is now able to use the concept banks from its eDiscovery to monitor their business practices on an ongoing basis. In this way, the company can identify and address transgressions before they turn into significant problems.
Businesses need to attract new customers; however, it is just as critical to retain the customers they already have. This requires excellent customer service, which can’t be delivered without identifying and valuing customer sentiment. To understand the customer experience, analysts try to read through all forms of customer communications, from suggestions to recommendations to complaints. In this Internet age, the volume and variety of communication can make this task near impossible. Analysts are overwhelmed by digital feedback from emails, online surveys, online discussion groups, blogs and more.
This explosion of data is mainly in the form of unstructured text, which adds to your challenge expotentially. These narratives can use very different language to express the same ideas and sentiments, thus making it difficult for readers to capture all the connections and produce real insights.
CONQ solves this problem with patented technologies. First, our groundbreaking software harmonizes and fuses disparate data, thus enabling the user to analyze the information in its natural state. There is no longer a need to extract, transform and load data—time-consuming and inefficient IT functions. Second, the analysts can apply their subject matter expertise by developing “Concepts” that are ideas, terms or phrases that capture the expression of customer sentiment.
CUSTOMER SENTIMENT CASE STUDY
CUSTOMER SENTIMENT CASE STUDY
CONQ’s data analytics helps a large financial services firm understand customer sentiment in order to improve its online experience.
For this bank, one of the five largest financial services companies in the United States, quality customer service is at the heart of business strategy. To this end, the bank’s e-Commerce division has a Customer Advocacy Group that seeks to identify and understand the “voice of the customer” and pass along critical feedback to individual product teams. The fast-growing popularity of online banking has focused the group on customers’ online experience. Knowing that it takes continual innovation to retain online customers, the bank views managing the customer experience as a key strategic initiative.
To provide better service, the bank focuses on customer suggestions, recommendations, and complaints. The amount of feedback has grown dramatically in this digital age, thus overwhelming the resources of the bank’s 10-person Customer Advocacy Group. The team receives more than 25,000 customer communications each month from varied sources such as emails, online surveys, and data recorded on customer relationship management (CRM) systems.
Not only do team members need to read through all the data, but the communications are in a free-form text format. As a result, it is very hard for the group to capture valuable information and produce constructive insights. They needed a software solution that can help uncover hidden trends, relationships, and patterns in unstructured data—and so they turned to CONQ.
The Jacksonville Police needed a way to quickly identify and extract relevant information from vast volumes of text, with the added difficulty that related Concepts might be expressed using different terms and spellings. Otherwise, investigators could spot the trends and make the connections necessary to solve the case.
With CONQ, the Customer Advocacy Group no longer relies on capturing information manually, which enables the team to perform advanced analytics, which it was previously unable to do. So much more was accomplished quickly and efficiently. In one case, the team was able to cut the time it took to read the monthly customer communications from one week to just one hour. This greater productivity allowed the group to focus more on value-added analysis and less on labor intensive reading to prepare the reports.
Moreover, the analysts used CONQ to easily identify top themes in customer communications, create concept banks from them, and thereby spot hard-to-find trends pertaining to product ideas and recommendations. CONQ has also allowed for an unprecedented holistic view of the customer experience by aggregating data from different sources to identify issues across the lines of business. The result is that product teams are now procuring precise quantitative analysis of customer feedback, an essential step in improving customer relations.
The Customer Advocacy Group is considering a proactive application of CONQ: using its unstructured analytics as an “early warning system” to quickly address developing problems in the customer experience. The group believes this capability could be a strategic differentiator in the competitive world of financial services.