# US 2026/0212417 A1 · Generative artificial intelligence for financial planning

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

- Number: US 2026/0212417 A1
- Document ID: US20260212417A1
- Status: Published application
- Kind: A1
- Application number: 19/032,700
- Filed: 2025-01-21
- Published: 2026-07-23
- Earliest claimed: 2025-01-21
- Applicants: Wells Fargo Bank, N.A.
- Inventors: Pawan Kumar Jha; Ganesh Lakshmanan Thondikulam
- Web page: https://patentgrep.com/patent/US20260212417A1

## CPC

- G06Q 40/06: Asset management; Financial planning or analysis
- G06Q 40/10: Tax strategies
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

An example computer system for financial planning can: receive, from third party data sources, financial data; generate, using a generative artificial intelligence model, personalized investment models for users based on the financial data and financial profiles of the users; evaluate, using adaptive machine learning techniques, risks associated with the personalized investment models for the users; generate personalized financial recommendations for the users based on the personalized investment models and evaluated risks, wherein the personalized financial recommendations include financial strategies; continuously update the personalized investment models and financial recommendations by causing the computer system to: detect drift in financial market patterns using online learning algorithms; automatically retrain the generative artificial intelligence model based on the drift; and adjust the financial recommendations based on the generative artificial intelligence model; and provide, through a user interface, the personalized financial recommendations that enable the users to simulate various financial scenarios and adjust investment strategies.

## Claims

### Claim 1

A computer system for financial planning, 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 third party data sources, real financial data through connectivity with multiple data streams;
generate synthetic financial data using a generative adversarial network architecture comprising a generator network that creates synthetic financial scenarios based on input data, and a discriminator network that evaluates the synthetic financial scenarios by distinguishing between the synthetic financial data and the real financial data, wherein the generator network and the discriminator network are trained iteratively through adversarial training such that the generator network learns to produce realistic synthetic financial scenarios;
validate the synthetic financial data through comparison with patterns from the real financial data to develop realistic market behaviors and statistical properties;
combine, upon validation, the synthetic financial data with the real financial data to create an augmented training dataset, wherein the synthetic financial data fills in historical gaps in the real financial data to allow for improved model training;
generate, using a generative artificial intelligence model trained with the augmented training dataset, personalized investment models for users based on the real financial data and financial profiles of the users;
evaluate, using adaptive machine learning techniques, risks associated with the personalized investment models for the users;
generate personalized financial recommendations for the users based on the personalized investment models and evaluated risks, wherein the personalized financial recommendations include financial strategies;
continuously update the personalized investment models and financial recommendations by causing the computer system to:
detect drift in financial market patterns using online learning algorithms;
automatically retrain the generative artificial intelligence model based on the drift; and
adjust the financial recommendations based on the generative artificial intelligence model; and
provide, through a user interface, the personalized financial recommendations that enable the users to simulate various financial scenarios and adjust investment strategies.

### Claim 2

(canceled)

### Claim 3 (depends on 1)

The computer system of claim 1, wherein, to evaluate risk, the instructions further cause the computer system to:
analyze short-term market volatility, medium-term economic cycles, and long-term demographic trends using multi-horizon predictive models;
evaluate personal risk factors including employment sector risks and family-specific circumstances; and
generate risk mitigation strategies.

### Claim 4 (depends on 1)

The computer system of claim 1, wherein the instructions further cause the computer system to generate:
a wizard that guides the users through defining financial objectives based on their unique circumstances;
a risk tolerance questionnaire that dynamically adapts based on user responses;
a simulator for testing various financial decisions; and
a portfolio rebalancing tool that provides real-time allocation recommendations.

### Claim 5 (depends on 1)

The computer system of claim 1, wherein, to generate the personalized financial recommendations, the instructions further cause the computer system to:
analyze retirement planning needs based on projected income requirements;
develop education funding strategies considering tax advantages;
create wealth preservation and transfer strategies across generations; and
provide asset allocation recommendations across different investment categories.

### Claim 6 (depends on 1)

The computer system of claim 1, wherein the instructions further cause the computer system to:
use the synthetic financial data for testing the financial strategies.

### Claim 7 (depends on 1)

The computer system of claim 1, wherein, to detect the drift, the instructions further cause the computer system to:
monitor prediction accuracy and model performance metrics;
identify changes in established market pattern correlations; and
initiate model retraining when significant pattern changes are detected.

### Claim 8 (depends on 1)

The computer system of claim 1, wherein the instructions further cause the computer system to:
provide visualization tools for displaying investment reports, risk analysis charts, and asset allocation data;
enable dynamic adjustment of financial strategies through interactive interfaces; and
generate real-time alerts based on personalized alert thresholds.

### Claim 9 (depends on 1)

The computer system of claim 1, wherein the adaptive machine learning techniques comprise:
reinforcement learning algorithms that optimize investment strategies based on simulated outcomes;
transfer learning techniques that adapt pre-trained models to individual user profiles; and
ensemble learning that combines multiple specialized models for risk assessment.

### Claim 10 (depends on 1)

The computer system of claim 1, wherein the instructions further cause the computer system to:
process both structured and unstructured data from the third party data sources;
normalize the real financial data across different asset classes and time scales; and
create a unified dataset for generating the personalized investment models.

### Claim 11

A method for financial planning, comprising:
receiving, from third party data sources, real financial data through connectivity with multiple data streams;
generating synthetic financial data using a generative adversarial network architecture comprising a generator network that creates synthetic financial scenarios based on input data, and a discriminator network that evaluates the synthetic financial scenarios by distinguishing between the synthetic financial data and the real financial data, wherein the generator network and the discriminator network are trained iteratively through adversarial training such that the generator network learns to produce realistic synthetic financial scenarios;
validating the synthetic financial data through comparison with patterns from the real financial data to develop realistic market behaviors and statistical properties;
combining, upon validation, the synthetic financial data with the real financial data to create an augmented training dataset, wherein the synthetic financial data fills in historical gaps in the real financial data to allow for improved model training;
generating, using a generative artificial intelligence model trained with the augmented training dataset, personalized investment models for users based on the real financial data and financial profiles of the users;
evaluating, using adaptive machine learning techniques, risks associated with the personalized investment models for the users;
generating personalized financial recommendations for the users based on the personalized investment models and evaluated risks, wherein the personalized financial recommendations include financial strategies;
continuously updating the personalized investment models and financial recommendations by:
detecting drift in financial market patterns using online learning algorithms;
automatically retraining the generative artificial intelligence model based on the drift; and
adjusting the financial recommendations based on the generative artificial intelligence model; and
providing, through a user interface, the personalized financial recommendations that enable the users to simulate various financial scenarios and adjust investment strategies.

### Claim 12

(canceled)

### Claim 13 (depends on 11)

The method of claim 11, further comprising analyzing short-term market volatility, medium-term economic cycles, and long-term demographic trends using multi-horizon predictive models;
evaluating personal risk factors including employment sector risks and family-specific circumstances; and
generating risk mitigation strategies.

### Claim 14 (depends on 11)

The method of claim 11, further comprising:
providing a wizard that guides the users through defining financial objectives based on their unique circumstances;
providing a risk tolerance questionnaire that dynamically adapts based on user responses;
providing a simulator for testing various financial decisions; and
providing a portfolio rebalancing tool that provides real-time allocation recommendations.

### Claim 15 (depends on 11)

The method of claim 11, further comprising:
analyzing retirement planning needs based on projected income requirements;
developing education funding strategies considering tax advantages;
creating wealth preservation and transfer strategies across generations; and
providing asset allocation recommendations across different investment categories.

### Claim 16 (depends on 11)

The method of claim 11, further comprising:
using the synthetic financial data for testing the financial strategies.

### Claim 17 (depends on 11)

The method of claim 11, further comprising:
monitoring prediction accuracy and model performance metrics;
identifying changes in established market pattern correlations; and
initiating model retraining when significant pattern changes are detected.

### Claim 18 (depends on 11)

The method of claim 11, further comprising:
providing visualization tools for displaying investment reports, risk analysis charts, and asset allocation data;
enabling dynamic adjustment of financial strategies through interactive interfaces; and
generating real-time alerts based on personalized alert thresholds.

### Claim 19 (depends on 11)

The method of claim 11, further comprising:
using reinforcement learning algorithms that optimize investment strategies based on simulated outcomes;
using learning techniques that adapt pre-trained models to individual user profiles; and
using ensemble learning that combines multiple specialized models for risk assessment.

### Claim 20 (depends on 11)

The method of claim 11, further comprising:
processing both structured and unstructured data from the third party data sources;
normalizing the real financial data across different asset classes and time scales; and
creating a unified dataset for generating the personalized investment models.
