Synthetic User Models for Iterative Research Testing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conducting user research with human users is time-consuming and expensive, making it difficult to test ad-hoc ideas and concepts iteratively and repetitively, and human users may provide biased or incomplete feedback.

Innovation Solution

Using machine learning models to create synthetic users that can emulate different human personas, allowing for the extraction of feedback and user insights without the need for human participation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If human users are used for user research, then feedback can be obtained, but the process becomes time-consuming and expensive

Engineering Contradiction:

Improvefeedback qualityVSAvoidresearch time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent creates synthetic users that are copies or simulations of human users, trained on human user data to replicate human behavior patterns, preferences, and decision-making processes. These synthetic user models can be deployed to conduct user research without requiring actual human participants, thereby reducing time and cost while maintaining feedback quality.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the mechanical system of human user participation (recruitment, scheduling, interviewing, analysis) with an automated computational system. Machine learning models process user research tasks algorithmically, eliminating the need for manual human involvement in the research process while preserving the ability to generate meaningful feedback.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If human users are used for user research, then real feedback can be obtained, but the cost increases significantly

Engineering Contradiction:

Improvefeedback authenticityVSAvoidresearch cost

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The synthetic users are trained on extensive human user data to replicate authentic human behavior, preferences, and responses. This copying approach allows the system to generate reliable, authentic-looking feedback without incurring the costs associated with recruiting and compensating actual human users.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system uses human user data that has already been collected from previous interactions to train the synthetic user models. This self-service approach leverages existing data resources rather than requiring new human participants, thereby reducing costs while maintaining feedback reliability through the use of authentic user behavior patterns.

Inventive Principle:

Principle #25Self-service

3Loss of information

If traditional user research methods are used, then comprehensive insights can be gathered, but the process lacks efficiency

Engineering Contradiction:

Improveinsight completenessVSAvoidresearch efficiency

Core Design Contradiction:

Loss of informationVSProductivity

Solution Approach 1:

The synthetic user models can operate continuously without interruption, running user research simulations 24/7 without the constraints of human availability. Multiple synthetic users can process different research scenarios simultaneously, enabling parallel execution of research tasks and significantly improving productivity while maintaining comprehensive insight gathering.

Inventive Principle:

Principle #20Continuity of useful action

Solution Approach 2:

The system segments the user research process into multiple independent synthetic user instances, each capable of handling specific research tasks. This segmentation allows for parallel processing of different research questions, user personas, or product features, thereby increasing overall research efficiency without sacrificing the comprehensiveness of insights gathered.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250165678A1Use a generative model to create synthetic users for testing and analysis

Publication Date: 2025.05.22 ROKU INC

AI summary

Conducting user research with human users is time consuming and expensive. Provided with quality training data that represents human users having different personae, models can be trained to offer responses that can emulate different human users. The models may offer synthetic users that can extract information from and respond to inputs to the models. The inputs can include prompts and responses can include answers to the prompts. In some cases, data about human users can be used to build different synthetic user memories, which may be used to generate different prompt chains corresponding to different users. The prompt chain can be used to prompt a model to respond based on the contextual information in the synthetic user memory.