LLM Event Description Rewriting for Fundraising Engagement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users often struggle to create compelling descriptions and titles for fundraising campaigns, leading to reduced reach and potential contributions due to lack of confidence in crafting effective messages and uncertainty about how to convey urgency or importance.

Innovation Solution

A system that uses large language models (LLMs) to modify user-specified event descriptions and titles by incorporating query templates, evaluating and ranking alternatives to optimize them for donor engagement, and iteratively refining the LLMs for improved accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If users manually craft descriptions and titles for fundraising campaigns, then they have control over the content, but they lack confidence and effectiveness in creating compelling messages

Engineering Contradiction:
Improveeffectiveness of campaign descriptionsVSAvoiduser confidence in crafting messages
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system enables users to benefit from professional content creation without manual effort by automatically generating optimized descriptions and titles using machine learning models, allowing users to simply provide basic campaign information while the system handles the complex content creation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual writing processes with automated machine learning-based content generation systems that use natural language processing to create compelling descriptions and titles, substituting human creative labor with AI-driven mechanisms

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If users manually create campaign content, then they can express their vision, but they uncertainty about how to convey urgency and importance

Engineering Contradiction:
Improveability to convey urgency and importanceVSAvoidcomplexity of message crafting
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system provides iterative feedback by generating multiple content variations and allowing users to review and select from suggested options, with the machine learning model continuously refining its output based on user preferences and campaign performance data

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The machine learning model acts as an intermediary between the user's basic campaign information and the final polished content, translating simple inputs into compelling messages while maintaining user vision and intent

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If fundraising campaigns use basic descriptions, then they are easy to create, but they have reduced reach and potential contributions

Engineering Contradiction:
Improvecampaign reach and contributionsVSAvoidease of content creation
Core Design Contradiction:
ProductivityVSEase of manufacture

Solution Approach 1:

The system automatically generates optimized content that maximizes campaign reach and contribution potential without requiring users to invest significant time in content creation, with users simply needing to provide basic campaign details

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The machine learning model analyzes and optimizes multiple content parameters including tone, structure, keywords, and formatting to generate descriptions and titles that are statistically more likely to achieve better campaign performance

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS20250384102A1Modifying content using machine learning
Publication Date: 2025.12.18 GOFUNDME INC
  • US20250384102A1 patent drawing
  • US20250384102A1 patent drawing
  • US20250384102A1 patent drawing

AI summary

Systems and methods provide for modifying a description for an event. A description for an event and a request for modifying the description for the event is received. A query template is selected based on the request for modifying the description of the event. A query is generated using a query template for a machine learning model. A modified description for the event is received from a machine learning model which is provided for display on a website.