US 12,468,718 B1
Integrated multi-channel conversational utility
Hannah Furr, Charlotte, NC (US)
Assigned to Bank of America Corporation, Charlotte, NC (US)
Filed by Bank of America Corporation, Charlotte, NC (US)
Filed on Dec. 2, 2024, as Appl. No. 18/965,664.
Int. Cl. G06F 16/20 (2019.01); G06F 16/2457 (2019.01); G06F 16/9538 (2019.01)
CPC G06F 16/24578 (2019.01) [G06F 16/9538 (2019.01)] 20 Claims
OG exemplary drawing
 
1. An integrated, multi-channel, conversational utility, said utility comprising a processor and computer-executed instructions that, when run on the processor, are configured to:
train a machine-learning (ML) model to associate input keywords with a set of result options, wherein:
said set of result options comprises results across at least two separate information sources housed within an entity hosting the utility; and
the at least two separate information sources comprise a general information center and a user-specific information center;
provide a single search icon on a display of a remote device of a user accessing an online portal of the entity;
receive, via the single search icon, a conversational input inquiry;
generate, via the ML model:
a first set of results comprising at least one general information resource result and at least one user-specific information result; and
a confidence score of a relevance of the first set of results to the input inquiry;
when the confidence score exceeds a predetermined threshold score, display the first set of results on the display of the remote device as a response to the input inquiry; and
when the confidence score fails to exceed the predetermined threshold score:
generate, via the ML model, a conversational follow-up question designed to clarify an intent of the input inquiry;
display the follow-up question on the display of the remote device;
receive a conversational input response to the follow-up question;
generate, via the ML model, and based on the input inquiry and the input response:
a second set of results comprising at least one general information resource result and at least one user-specific information result; and
a second confidence score of a relevance of the second set of results to the input inquiry; and
when the second confidence score exceeds the predetermined threshold score, display the second set of results on the display of the remote device as the response to the input inquiry.