An AI-guided framework for making TRIZ and C–K more accessible

Abstract

TRIZ, as one of the well-known structured creative problem-solving methods, offers structured ways to approach problem-solving, but in practice it can be difficult to use. Its tool selection steps often require experience, which means many practitioners, especially beginners, find it challenging to apply. Issues such as unfamiliarity with the tools, gaps in relevant knowledge, and the tendency to fall back on familiar thinking patterns can make it hard to benefit fully from this method. Because of this, the strengths of TRIZ are not always reflected in everyday innovation work. This research explores how the developed framework utilizes artificial intelligence, particularly large language models like ChatGPT, and eases these difficulties by guiding users through TRIZ and, later, C–K theory in a more structured and manageable way. Instead of asking the AI for direct solutions, which often leads to generic or repetitive ideas, the proposed approach uses a series of focused questions based on the logic of TRIZ or C–K. These questions help users move step by step from defining the problem more clearly, recognizing contradictions, gathering relevant knowledge, and finally evaluating possible directions. This interaction allows users to stay in control of the process, adjusting or discarding AI output when needed, and helps avoid the homogenization often seen in unstructured LLM outputs. The four studies included in this thesis apply this idea in different contexts, from improving the design of non-slip footwear to supporting concept development for compact sanitation systems and exploring solutions related to driver safety. Across these cases, the findings proof that the suggested framework could make it easier to work with TRIZ and C–K principles, especially for users who are not deeply familiar with the methods. Users also are guided to approach complex problems more systematically and are less likely to get stuck in habitual thinking patterns. Overall, the work shows that application of AI in the framework can play a practical supporting role. By helping clarify steps, surface relevant information, and structure the reasoning process, the proposed frameworks make TRIZ and C–K more accessible to a wider range of users.

Date	17 Jun 2026
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Mickaël Gardoni (Supervisor)