US 11,055,772 B1
Instant lending decisions
Siddharth Ram, Menlo Park, CA (US); Richard N. Preece, San Diego, CA (US); Joseph Timothy Callinan, Jr., Campbell, CA (US); Kathy Tsitovich, Mountain View, CA (US); and Eva Diane Chang, Mountain View, CA (US)
Assigned to Intuit Inc., Mountain View, CA (US)
Filed by Siddharth Ram, Menlo Park, CA (US); Richard N. Preece, San Diego, CA (US); Joseph Timothy Callinan, Jr., Campbell, CA (US); Kathy Tsitovich, Mountain View, CA (US); and Eva Diane Chang, Mountain View, CA (US)
Filed on Nov. 21, 2018, as Appl. No. 16/198,599.
Application 16/198,599 is a continuation of application No. 13/956,281, filed on Jul. 31, 2013, abandoned.
Int. Cl. G06Q 40/02 (2012.01)
CPC G06Q 40/025 (2013.01) 23 Claims
OG exemplary drawing
 
1. A method, comprising:
obtaining user entered data from a business management application (BMA) for a plurality of business entities, wherein the user entered data is entered by the plurality of business entities and represents business activities performed by the plurality of business entities;
obtaining usage statistics from the BMA for the plurality of business entities, wherein the usage statistics comprise login statistics regarding a plurality of logins to the BMA by each of the plurality of business entities;
analyzing, by a machine learning algorithm, at least the user entered data and the usage statistics to generate a plurality of risk profiles for the plurality of business entities, wherein the plurality of risk profiles represent probabilities of the plurality of business entities defaulting on a loan;
providing, over a computer network, the plurality of risk profiles to a computing device of a first lender, wherein the computing device of the first lender executes a plurality of lending decisions with respect to the plurality of business entities based on the plurality of risk profiles;
training the machine learning algorithm by iteratively adjusting, by a computer processor, adjusted matching parameters of the machine learning algorithm to increase a correlation between approval statistics of the plurality of lending decisions and the plurality of risk profiles, wherein iteratively adjusting continues until reaching a threshold correlation between the approval statistics and the plurality of lending decisions and the plurality of risk profiles, wherein training generates an updated machine learning algorithm;
obtaining, by an application service provider that provides access to the BMA to the plurality of business entities, a number of logins to the BMA by a business entity from among the plurality of business entities; and
updating, over the computer network and in real time with obtaining the number of logins to the BMA, a risk score of a risk profile for the business entity, wherein the risk score of the risk profile for the business entity is updated using the number of logins to the BMA as inputs to the updated machine learning algorithm.