Customer-Specific Text Prediction With Dynamic Data Variables

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing text personalization methods, such as templates and digital assistant responses, are static and fail to adapt dynamically to changing customer preferences, making manual composition of personalized text messages labor-intensive and inefficient.

Innovation Solution

A text prediction system that uses customer-specific models to suggest personalized text for insertion at a cursor position in a text message, incorporating data variables replaceable with customer-specific values, and provides continuous type-ahead suggestions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If templates are used for text personalization, then some level of personalization is achieved, but the templates are static and cannot dynamically adapt to changing customer preferences

Engineering Contradiction:

Improveadaptability to customer preferencesVSAvoidstatic nature of templates

Core Design Contradiction:

Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The patent implements dynamic text generation by training machine learning models on customer-specific data, enabling the system to adapt text suggestions to changing customer preferences over time. The models are continuously updated with new data, allowing the text generation to evolve dynamically rather than relying on static templates.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes parameters by using customer-specific training data to adjust model behavior for different customers. The model parameters are customized per customer based on their communication patterns, preferences, and historical interactions, enabling dynamic adaptation without requiring static template configurations.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If manual personalization is performed, then high level of personalization is achieved, but the process is labor intensive and time consuming

Engineering Contradiction:

Improvelevel of personalizationVSAvoidcomposing efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The system provides self-service text generation by automatically analyzing customer data and generating personalized text suggestions without requiring manual composition. The machine learning models autonomously create customized text based on customer profiles, eliminating the need for labor-intensive manual personalization while maintaining high personalization quality.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical process of manual text composition with an automated machine learning-based system. The manual writing process is substituted by computational models that generate personalized text automatically, dramatically improving productivity while maintaining or enhancing personalization quality.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Adaptability or versatility

If customer-specific models are trained and used, then text personalization is dynamically adapted to customer preferences, but the model training and data processing complexity increases

Engineering Contradiction:

Improvedynamic personalization capabilityVSAvoidmodel training complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the customer base into individual customer profiles, each with their own trained model. By dividing the overall system into customer-specific modules, the complexity is distributed and managed more effectively. Each segment (customer) receives customized attention through dedicated modeling, enabling dynamic personalization without overwhelming system complexity.

Inventive Principle:

Principle #1Segmentation

4Adaptability or versatility

If static templates are used, then processing is simpler, but the ability to provide personalization is limited

Engineering Contradiction:

Improvepersonalization capabilityVSAvoidprocessing complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal machine learning framework that can handle multiple customer profiles and generate personalized text for various communication contexts. The same underlying model architecture serves multiple customers with different preferences, providing universal personalization capability without requiring separate complex systems for each customer.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS12456010B2Predicting text and data variables using customer-specific models

Publication Date: 2025.10.28 ORACLE INT CORP

AI summary

A text prediction system for providing personalized text suggestions for different customers is disclosed. For a text message being composed via an application for an intended recipient, the system receives a portion of the text message based upon a current cursor position in the text message. The system identifies a customer associated with the intended recipient and identifies a customer-specific model for the customer. The system predicts text to be suggested for type-ahead insertion at the current cursor position using the customer-specific model and based upon the portion of the text message. The suggested text may include one or more data variables that are replaceable with values determined from data stored for the customer. The data variables provide further customization to the suggested text by presenting suggestions are more personalized for the recipient of the text message. The system causes the suggested text to be output via the application.