# TRIZ-GPT: An LLM-augmented method for problem-solving

arXiv: https://arxiv.org/abs/2408.05897 · HTML: https://mlsanity.com/paper/2408.05897/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Aug 2024.

## Built on by (most cited first, 3 of 3)

- [TRIZ Agents: A Multi-Agent LLM Approach for TRIZ-Based Innovation](https://mlsanity.com/paper/2506.18783/) (2025)
- [AutoSAM: an Agentic Framework for Automating Input File Generation for the SAM Code with Multi-Modal Retrieval-Augmented Generation](https://mlsanity.com/paper/2603.24736/) (2026)
- [Enhancing Creativity in 3D Generative Design via a TRIZ-Inspired Text-to-CAD Framework](https://mlsanity.com/paper/2606.21378/) (2026)

## Abstract (arXiv)

TRIZ, the Theory of Inventive Problem Solving, is derived from a comprehensive analysis of patents across various domains, offering a framework and practical tools for problem-solving. Despite its potential to foster innovative solutions, the complexity and abstractness of TRIZ methodology often make its acquisition and application challenging. This often requires users to have a deep understanding of the theory, as well as substantial practical experience and knowledge across various disciplines. The advent of Large Language Models (LLMs) presents an opportunity to address these challenges by leveraging their extensive knowledge bases and reasoning capabilities for innovative solution generation within TRIZ-based problem-solving process. This study explores and evaluates the application of LLMs within the TRIZ-based problem-solving process. The construction of TRIZ case collections establishes a solid empirical foundation for our experiments and offers valuable resources to the TRIZ community. A specifically designed workflow, utilizing step-by-step reasoning and evaluation-validated prompt strategies, effectively transforms concrete problems into TRIZ problems and finally generates inventive solutions. Finally, we present a case study in mechanical engineering field that highlights the practical application of this LLM-augmented method. It showcases GPT-4's ability to generate solutions that closely resonate with original solutions and suggests more implementation mechanisms.