US 12,243,107 B2
Usage estimation systems and methods for risk association adjustments
Brandon A. Banks, West Hartford, CT (US); and Kenneth J. Zygiel, Harwinton, CT (US)
Assigned to HARTFORD FIRE INSURANCE COMPANY, Hartford, CT (US)
Filed by HARTFORD FIRE INSURANCE COMPANY, Hartford, CT (US)
Filed on Sep. 19, 2023, as Appl. No. 18/470,014.
Application 18/470,014 is a continuation of application No. 17/066,803, filed on Oct. 9, 2020, granted, now 11,798,093.
Prior Publication US 2024/0005412 A1, Jan. 4, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 40/08 (2012.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01); G16Y 40/10 (2020.01)
CPC G06Q 40/08 (2013.01) [G06N 5/04 (2013.01); G06N 20/00 (2019.01); G16Y 40/10 (2020.01)] 20 Claims
OG exemplary drawing
 
1. An estimated usage risk relationship management system implemented via a back-end application computer server, comprising:
(a) a risk relationship data store that contains electronic records, each electronic record representing a risk relationship between an enterprise and a risk relationship provider, and including, for each risk relationship, an electronic record identifier and a predicted usage attribute value;
(b) a current usage data source associated with the enterprise;
(c) the back-end application computer server, coupled to the risk relationship data store and the current usage data source, including:
a computer processor, and
a computer memory, coupled to the computer processor, storing instructions that, when executed by the computer processor, cause the back-end application computer server to:
(i) receive, from the current usage data source, current Internet of Things (IoT) information for the enterprise for a resource,
(ii) receive third-party data and historical data for the resource;
(iii) based on the current IoT information for the resource, the received third-party data for the resource and the received historical data for the resource, infer, via a trained machine learning model, a likely actual current usage for the enterprise for the resource;
(iv) compare the likely actual current usage with the predicted usage attribute value to determine a risk difference result;
(v) automatically adjust a risk relationship parameter based on the risk difference result and a fluctuation sensitivity level;
(vi) re-train the trained machine learning model in response to receipt of the adjusted risk relationship parameter; and
(d) a communication port coupled to the back-end application computer server to facilitate a transmission of data to a remote device to support an interactive display, including an indication of the adjusted risk relationship parameter, via a distributed communication network.