How does prompting affect innovation and creative processes?
Four main phases of the research methodology
Four main phases of the research methodology
Methodology flow diagram illustrating the four main phases of the study
Research Question & Purpose
My research looked at how using structured prompts can improve idea generation when collaborating with AI. Specifically, I explored whether clear, structured prompting techniques lead to better ideas and a more satisfying user experience during AI-assisted brainstorming sessions.
This topic matters especially for teams or organizations that want to get the most out of AI tools in their innovation processes, making sure they aren't just integrating technology, but doing it effectively.
Methodology
I employed a mixed-methods approach combining standardized questionnaires with creative ideation tasks:
Participant Groups
67 participants were divided into two groups through video training: one group learned structured prompting techniques, while the other received general AI capabilities training.
Ideation Task
Participants were asked to generate product ideas for university students using ChatGPT 4o, applying their respective training approaches.
Evaluation Criteria
Ideas were evaluated on multiple dimensions: relevance, feasibility, market potential, innovation, and scalability.
Data Collection
Quantitative data from standardized assessments was complemented by qualitative analysis of the ideas generated and participant feedback.
Screenshot from the prompting techniques training video
Screenshot from the prompting techniques training video
Example of the ChatGPT interface during the ideation task
Example of the ChatGPT interface during the ideation task
Key Findings
The study revealed several significant findings about the impact of structured prompting on AI-assisted ideation:
Higher Feasibility Scores
Participants trained in prompting achieved significantly higher feasibility scores for their generated ideas, indicating more practical and implementable solutions.
Greater User Satisfaction
Prompting-trained participants reported higher satisfaction with their outputs, suggesting that structured communication with AI increases confidence in results.
Similarity Trade-off
Prompting-trained ideas showed higher similarity, suggesting a potential trade-off between quality and originality when using structured approaches.
Technical Affinity Impact
Participants with lower technical affinity benefited more from structured prompting approaches, indicating that prompting training may have equalizing effects.
Comparison of feasibility scores between prompting-trained and control groups
Comparison of feasibility scores between prompting-trained and control groups
Self-reported satisfaction scores across participant groups
Self-reported satisfaction scores across participant groups
Cosine similarity comparison showing idea diversity between groups
Cosine similarity comparison showing idea diversity between groups
Implications
The research findings have several important implications for organizations and individuals working with AI tools:
Enhanced User Confidence
Structured prompting enhances user confidence and satisfaction with AI-generated outputs, potentially leading to greater adoption of AI tools in creative processes.
Training Recommendations
Organizations should consider implementing prompting training, especially for users less comfortable with AI technology, to maximize the value of AI-assisted ideation.
Balancing Structure and Creativity
Future development should focus on balancing structured approaches with techniques to enhance creative diversity, addressing the potential trade-off between quality and originality.
Democratizing AI Benefits
Prompting training may help democratize the benefits of AI-assisted ideation, making advanced technologies more accessible to users with varying levels of technical expertise.
Download Full Thesis
Interested in the complete research? Download the full thesis document for detailed methodology, results, and analysis.
Download Thesis PDFAcknowledgments
I would like to express my gratitude to everyone who contributed to this research:
- Prof. Dr. Christian Lüthje for his invaluable guidance and support
- Prof. Dr. Cornelius Herstatt for his insightful feedback and direction
- All research participants who contributed their time and creativity
- The Institute of Innovation Marketing for providing resources and support