CONFIDENCE INDICATOR

CONFIDENCE INDICATOR TECHNIQUE OVERVIEW

The Confidence Indicator technique is designed to measure a level of trust or confidence in a response to a given question relying on behavioral analysis and cognitive science. Potentially within an existing application process, consumers will be presented questions that (by design) may invoke an emotional response, which can be detected in fine motor movements and how the consumer responds to the question. This information is interpreted based on extensive neuro-cognitive research to provide an indication of how confident an organization can be in whether responses to these questions were truthful or deceptive. Examples of pertinent or risk-relevant questions include asking credit applicants if they recently lost employment or expect to lose employment in the next few months, and asking whether all the information submitted for an auto insurance claim was truthful and accurate.

Key considerations when implementing or buying this functionality include:

• Do you already have an application process or questionnaire where additional questions could be added? Examples include credit card and loan applications, new account applications, as well as when consumers are submitting disputes or claims (such as insurance claims, chargebacks and returns).

• How will you make use of any confidence indication score or results in your rules or risk modeling engine? Do you plan to make it available for use with agent reviews or claim investigations?

• Think about what types of questions you may already be asking or should be asking that are pertinent to understanding the risk of your applicant or customer and ensure these can be supported by the Confidence Indicator vendor. There may be additional costs associated with creating and testing custom questions.

• Is this is a technique your organization should leverage for all or only high risk applications, claims or consumers?

HOW DOES IT WORK?

There are two primary components to how Confidence Indicator services work: the underlying science and research, and the technology to measure and analyze fine motor movement and behavioral nuances. The technology for measuring and detecting the fine motor movements and other activity related to how someone uses or navigates a web page on a computer or mobile device is very similar to what behavioral biometric and behavioral analytic providers are doing. Looking at how someone holds a mobile device, their pressure and speed on the touch screen, their speed and movement of a mouse – these are all things both types of vendors may look at or measure. Whereas behavioral biometric and monitoring vendors are looking to identify return users based on these physical and fine motor movement measurements, Confidence Indicator vendors are able to detect stress and other signals of potential deceit in response to specific stimuli (such as carefully crafted questions that provide insight about certain risks).

The Confidence Indicator approach relies on significant scientific research and experiments measuring the relationship between neurological states and how such changes influence fine motor movements and other indications of distress or emotional states that are captured via a computer or mobile device. Within a questionnaire or application process the consumer must be asked baseline questions that would not trigger any emotional response; then, when presented a question that may produce a neurological change, there will be measurable behavioral indications. Based on years of research and controlled experiments, fine motor movement and other behavioral signals indicative of different neurological responses have been identified, and when applied to a variety of contexts and questions, can show whether or not an organization can be confident in the legitimacy of an end user’s response.

In addition to understanding how this technique works it is important to understand how Confidence Indicators as a risk management technique are applied or are actually used. 

There has to be a baseline behavioral biometric measurement, some emotional stimuli, and post-stimuli behavioral biometric measurements for comparison. This stimuli and baseline are typically in the form of questions asked directly to the end user, and the design of these questions is critical in the overall value these services will provide. Confidence Indicator vendors should have pre-set questions for certain industries and applications, and may be able to support custom applications and collaboratively design new questions pertinent to an organization’s specific types of risk. There may be an iterative tuning process to develop the strongest set of baseline and risk relevant questions.

Confidence Indicator services are most commonly utilized by organizations that offer credit issuance or any type of loan application (personal, automobile, mortgage) as well as in the insurance industry to validate policyholder claims. In one context these services can be used to assess the credit or default risk of a new loan or credit applicant, and in the other, Confidence Indicator services can assess the fraud risk associated with an auto or property insurance claim (to flag false or inflated claims). In both applications this is accomplished by presenting specific and carefully crafted questions that provide insight and information relevant to the application or claim, as well as an indication of how confident we can be that the information provided is genuine. The questions themselves are varied and diverse, contingent on the organization and their industry. These questions can inquire about all types of information, but a major source of value is being able to ask questions about the future or future expectations, or information that otherwise cannot be validated or confirmed. Confidence Indicator services can provide a unique level of foresight and indication of truthfulness or deceit regarding many types of information that previously could not be confirmed.

HOW DO YOU USE THE RESULTS?

Results from Confidence Indicator providers can be provided in one of multiple ways to include an overall confidence indicator score for a full application or questionnaire, a confidence indicator score for each risk-relevant question, or providing detailed user results that would be used in manual review or claim investigations. When provided as a confidence indicator score, either at the full application or individual question level, this is easily used as an input within a rules engine or risk modeling platform. Applicants or claims that flag certain questions or have an application score above a certain threshold can be held for additional screening.

Providers may offer additional, more detailed signals that can be helpful in performing credit risk reviews or claim investigations. This can include detailed messages or indicators on what type of behavior a user displayed (such as answer switching) or even show full playbacks of how that person went through and completed the questionnaire or application.