Artificial Intelligence and What Wealthy Clients “Really” Want

This week Cetera Financial Group, one of the country’s largest independent broker-dealers, announced a pilot program for a new client profiling tool called Decipher. In the announcement, Decipher was described as a “unique, interactive technology that transforms the client discovery process, providing enhanced insights into client needs and concerns.”

Decipher works by combining facial recognition with artificial intelligence. Clients log in at a computer with a camera, answer a series of questions and view preset situational videos. The Decipher program maps clients’ reactions against a database of facial expressions signifying emotions. It then provides a readout into client emotions and behavior patterns regarding financial goals, decisions and concerns.

The use of biometrics, including facial recognition, has become increasingly common among financial institutions. But applications have largely been confined to relatively simple functions like managing account access and making sure account information is secure from unwanted intrusion. Decipher dramatically expands the role of these technologies, apparently using them to determine how clients “really” feel about their financial goals and how much risk they can tolerate.

The obvious concern with Decipher is whether the results can be trusted to be accurate and how accuracy would ultimately be measured. A scenario, for example, could be envisioned where a client’s verbal characterization of their risk level is very different from what the Decipher algorithm says it is. In such instances, does the advisor side with the man or the machine?

Cetera claims that “the new Decipher offering supports the DOL Fiduciary Rule’s key objectives and Cetera’s Best Interest Advice Model.” Consider again the scenario where the client’s self-perception differs from Decipher’s assessment. In such a case, would the advisor be in violation of the DOL Fiduciary Rule by listening to what the client says about themselves¬†rather than Decipher? And on what grounds would this decision be defended?

It will be interesting to see how Decipher is received and used in the field and to what degree investors accept the judgment of a machine particularly when it contradicts their own “conscious” beliefs about themselves.