Personalized recommendations
An example computer system for determining a recommendation can include: one or more processors; and non-transitory computer-readable storage media encoding instructions which, when executed by the one or more processors, causes the computer system to: receive a signal associated with an action or an inaction by a user; generate a machine learning model using the signal; process the machine learning model through a rules engine to generate the recommendation; and present the recommendation to the user through a non-traditional communication channel.
1 . A computer system for determining a recommendation, comprising:
one or more processors; and
non-transitory computer-readable storage media encoding instructions which, when executed by the one or more processors, cause the computer system to:
receive, from one of a plurality of electronic sources, a signal associated with an action or an inaction by a user;
generate a machine learning model using the signal, including to:
cluster the signal from the one of the plurality of electronic sources with further signals to determine patterns of user behaviors, wherein the further signals include interactions with smart devices in proximity-enabled environments to determine which of a plurality of non-traditional communication channels will be more conducive to preferred interactions;
train the machine learning model using the signal and the further signals as clustered to identify user preferences; and
dynamically update the machine learning model based on the patterns of user behaviors and the user preferences relating to an account portal, wherein the machine learning model continuously processes information about the patterns of user behaviors and the user preferences captured over multiple sessions, from the signal and the further signals, associated with the account portal to refine the machine learning model and provide more personalized recommendations, wherein the machine learning model is unique to the user and dynamically updates based on changes to account information for the user captured on the account portal over the multiple sessions;
process the machine learning model through a rules engine to generate the recommendation, wherein the rules engine comprises interrelated rules programmed to identify the personalized recommendations based upon the signal and the further signals, wherein the interrelated rules include qualifying rules having thresholds associated with application of each rule and priorities defining a hierarchy for presenting the personalized recommendations to the user, and wherein the qualifying rules are satisfied before triggering generation of the recommendation;
detect, using at least one device, that the user has entered a proximity-enabled environment of the proximity-enabled environments integrated with the smart devices;
select, using the machine learning model, a non-traditional communication channel from the plurality of non-traditional communication channels based upon a proximity of the user provided by the at least one device in the proximity-enabled environment, wherein the non-traditional communication channel is more conducive to the preferred interactions; and
present the recommendation to the user through the non-traditional communication channel on one of the smart devices, wherein the recommendation is presented in audio form through a voice assistant of the one of the smart devices based on detection of the user entering the proximity-enabled environment.
2 . The computer system of claim 1 , wherein the non-traditional communication channel includes one or more of: smart assistant, push notification, chatbot, audio call, smart vehicles, smart appliances, and an Internet of Things device.
3 . The computer system of claim 1 , wherein the recommendation is presented by a chatbot or an agent chat, and wherein the further signals from the user in response to the recommendation are sent to the rules engine.
4 . The computer system of claim 1 , wherein the signal is received from a third-party source.
5 . The computer system of claim 1 , wherein the rules engine makes a decision based upon a trigger captured by the signal.
6 . The computer system of claim 1 , comprising further instructions which, when executed by the one or more processors, cause the computer system to generate the machine learning model using artificial intelligence.
7 . The computer system of claim 1 , comprising further instructions which, when executed by the one or more processors, cause the computer system to send a notification to a third-party application.
8 . The computer system of claim 1 , wherein the signal is associated with home lending.
9 . The computer system of claim 8 , comprising further instructions which, when executed by the one or more processors, cause the computer system to generate a personal recommendation associated with an account for the home lending.
10 . The computer system of claim 8 , comprising further instructions which, when executed by the one or more processors, cause the computer system to capture a pattern of the user associated with an account for the home lending.
11 . A computer-implemented method for determining a recommendation, the method comprising:
receiving, from one of a plurality of electronic sources, a signal associated with an action or an inaction by a user;
generating a machine learning model using the signal, including:
clustering the signal from the one of the plurality of electronic sources with further signals to determine patterns of user behaviors, wherein the further signals include interactions with smart devices in proximity-enabled environments to determine which of a plurality of non-traditional communication channels will be more conducive to preferred interactions;
training the machine learning model using the signal and the further signals as clustered to identify user preferences; and
dynamically updating the machine learning model based on the patterns of user behaviors and the user preferences relating to an account portal, wherein the machine learning model continuously processes information about the patterns of user behaviors and the user preferences captured over multiple sessions, from the signal and the further signals, associated with the account portal to refine the machine learning model and provide more personalized recommendations, wherein the machine learning model is unique to the user and dynamically updates based on changes to account information for the user captured on the account portal over the multiple sessions;
processing the machine learning model through a rules engine to generate the recommendation, wherein the rules engine comprises interrelated rules programmed to identify the personalized recommendations based upon the signal and the further signals, wherein the interrelated rules include qualifying rules having thresholds associated with application of each rule and priorities defining a hierarchy for presenting the personalized recommendations to the user, and wherein the qualifying rules are satisfied before triggering generation of the recommendation;
detecting, using at least one device, that the user has entered a proximity-enabled environment of the proximity-enabled environments integrated with the smart devices;
selecting, using the machine learning model, a non-traditional communication channel from the plurality of non-traditional communication channels based upon a proximity of the user provided by the at least one device in the proximity-enabled environment, wherein the non-traditional communication channel is more conducive to the preferred interactions; and
presenting the recommendation to the user through the non-traditional communication channel on one of the smart devices, wherein the recommendation is presented in audio form through a voice assistant of the one of the smart devices based on detection of the user entering the proximity-enabled environment.
12 . The method of claim 11 , wherein the non-traditional communication channel includes one or more of: smart assistant, push notification, chatbot, audio call, smart vehicles, smart appliances, and an Internet of Things device.
13 . The method of claim 11 , wherein the recommendation is presented by a chatbot or an agent chat, and wherein the further signals from the user in response to the recommendation are sent to the rules engine.
14 . The method of claim 11 , wherein the signal is received from a third-party source.
15 . The method of claim 11 , wherein the rules engine makes a decision based upon a trigger captured by the signal.
16 . The method of claim 11 , further comprising generating the machine learning model using artificial intelligence.
17 . The method of claim 11 , further comprising sending a notification to a third-party application.
18 . The method of claim 11 , wherein the signal is associated with home lending.
19 . The method of claim 18 , further comprising generating a personal recommendation associated with an account for the home lending.
20 . The method of claim 18 , further comprising capturing a pattern of the user associated with an account for the home lending.