# US 2025/0148360 A1 · System and method for pre-loading content in a web environment

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0148360 A1
- Document ID: US20250148360A1
- Status: Published application
- Kind: A1
- Application number: 18/504,771
- Filed: 2023-11-08
- Published: 2025-05-08
- Earliest claimed: 2023-11-08
- Assignees: American Express Travel Related Services Company, Inc.
- Applicants: American Express Travel Related Services Company, Inc.
- Inventors: Rick BALLMAN; Shanmuga KRISHNAMURTHY
- Web page: https://patentgrep.com/patent/US20250148360A1

## CPC

- G06N 20/00: Machine learning
- H04L 67/535: {Tracking the activity of the user (network monitoring arrangements H04L43/00; recording of computer activity G06F11/34)}
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Disclosed herein are system, method, and computer program product embodiments for providing predictive interface generation. As a user interacts with a user interface, information relating to the user, including the user activity within the user interface and/or historical data relating to the user is provided to a machine learning model. The machine learning model uses artificial intelligence and the received user data to predict a likely next action that the user will take. In some instances, multiple next actions are determined, each with a corresponding confidence score. A most likely subset of those results are then provided to a preloading system that retrieves the data from a database that is needed to provide the response to those actions. This data is then preloaded in memory for immediate access, in the event that one of the predicted actions is actually carried out by the user.

## Claims

### Claim 1

A web prediction system, comprising:
a memory that stores data associated with a user;
an interface configured to interact with the user during a user session; and
one or more processors communicatively coupled to the memory and the interface, wherein the one or more processors are configured to:
receive, during the user session, user activity data from the interface;
predict a likely next action that the user will take by applying the received user activity data to a machine learning model trained to identify one or more web interface interactions corresponding to user activity data;
fetch web interface data corresponding to the likely next action; and
load the fetched web interface data into the memory prior to detecting an actual next action of the user.

### Claim 2 (depends on 1)

The web prediction system of claim 1, wherein the user activity data includes behavioral data of the user.

### Claim 3 (depends on 1)

The web prediction system of claim 1, wherein the user activity data includes at least one of historical and environmental data relating to the user.

### Claim 4 (depends on 1)

The web prediction system of claim 1, wherein the one or more processors are configured to predict a plurality of likely next actions, each ranked according to a confidence score.

### Claim 5 (depends on 1)

The web prediction system of claim 1, wherein the user activity data includes one of behavioral data or historical data relating to the user depending on a phase of the user session.

### Claim 6 (depends on 1)

The web prediction system of claim 1, wherein the user activity includes the historical data during an initial phase of the user session, and includes the behavioral data during an ongoing phase of the user session.

### Claim 7 (depends on 1)

The web prediction system of claim 1, wherein the one or more processors are further configured to:
detect the actual next action of the user;
determine whether the detected actual next action of the user corresponds to a likely next action loaded into the memory; and
generate the interface or perform a new fetch depending on the determining.

### Claim 8

A method for predictive preloading of a web environment during a user session, comprising:
receiving, during the user session, user activity data of the user;
predicting a likely next action that the user will take within the web environment by applying the received user activity data to a machine learning model trained to identify one or more web interface interactions corresponding to the user activity data;
fetching web interface data corresponding to the likely next action; and
loading the fetched web interface data into memory prior to detecting an actual next action of the user.

### Claim 9 (depends on 8)

The method of claim 8, wherein the user activity data includes behavioral data of the user.

### Claim 10 (depends on 8)

The method of claim 8, wherein the user activity data includes at least one of historical and environmental data of the user.

### Claim 11 (depends on 8)

The method of claim 8, further comprising predicting a plurality of likely next actions, each ranked according to a confidence score.

### Claim 12 (depends on 8)

The method of claim 8, wherein the user activity data includes one of behavioral data or historical data relating to the user depending on a phase of the user session.

### Claim 13 (depends on 8)

The method of claim 8, wherein the user activity data includes the historical data during an initial phase of the user session, and includes the behavioral data during an ongoing phase of the user session.

### Claim 14 (depends on 8)

The method of claim 8, further comprising:
detecting the actual next action of the user;
determining whether the detected actual next action of the user corresponds to a likely next action loaded into the memory; and
generating the interface or performing a new fetch depending on the determining.

### Claim 15

A non-transitory computer-readable storage medium storing instructions thereon that when read executed by one or more processors cause the one or more processors to execute functions, comprising:
receiving a user identity of a user;
retrieving historical and environmental data associated with the user based on the user identity;
calculating an initial next action prediction by applying the historical and environmental data to a machine learning model trained to identify one or more web interface interactions corresponding to historical and environmental data;
fetching initial web interface data corresponding to the initial next action;
preloading the fetched initial web interface data into memory prior to detecting an actual next action of the user;
receiving user activity data of the user;
calculating a subsequent next action prediction based on the user activity data;
fetching subsequent data needed to carry out the subsequent next action; and
preloading the fetched subsequent data into the memory prior to determining an actual next action of the user.

### Claim 16 (depends on 15)

The non-transitory computer-readable storage medium of claim 15, wherein the user activity data includes behavioral data of the user.

### Claim 17 (depends on 15)

The non-transitory computer-readable storage medium of claim 15, wherein each of the initial next action prediction and the subsequent next action prediction predict a plurality of likely next actions, each ranked according to a confidence score.

### Claim 18 (depends on 15)

The non-transitory computer-readable storage medium of claim 15, wherein the initial next action prediction or the subsequent next action is performed based on a phase of a user session.

### Claim 19 (depends on 18)

The non-transitory computer-readable storage medium of claim 18, wherein the initial next action prediction is performed during an initial phase of the user session and wherein the subsequent next action prediction is performed during an ongoing phase of the user session.

### Claim 20 (depends on 15)

The non-transitory computer-readable storage medium of claim 15, wherein the functions further comprise:
detecting the actual next action of the user;
determining that the actual next action corresponds to the predicted next action; and
generating a user interface from the data preloaded into the memory.