InnovationManagementMarketingEconomics and Finance
Sign In Subscribe

Management

Your new innovative team member: Creative AI? 

By Dr. Christian Byrge | January 8, 2026
Business Insights | Innovation | Management

Abstract

In a series of studies, AI has proven to be more innovative during creative problem-solving than most humans. How can we implement this enormous innovative resource for developing our next competitive advantages?

Photo by Pavel Danilyuk/Pexels

Introduction

Creativity is the engine for innovation. It naturally fosters new ideas within the organization: ideas for improving work processes, ideas for developing stronger relationships with customers and stakeholders, ideas for enhancing products and services, ideas for cost reduction, ideas for exploring new sales channels, and ideas for everything related to everyday business operations. Consequently, these ideas replace declining and less effective ones, leading to new competitive advantages (Drejer et al., 2019).

To solve complicated problems, we require a high quantity of data of high quality. We also need to analyse this data to predict which known solution might best address the problem. To solve complex problems, data and analysis may also be necessary. However, due to the uncertain and unpredictable nature of complex problems, data and analysis alone are insufficient. Complex problems demand new solutions, making the key ingredients for their resolution a high quantity of high-quality ideas. We also need to explore these ideas to develop a new solution for addressing these problems (Byrge, 2023).

The more great ideas we generate, the more opportunities we can choose from, providing a stronger foundation for decision-making when tackling complex problems. Consequently, creative thinking assists in making better decisions when addressing complex problems.

This paper examines key creative tasks where teams can utilize AI in their creative problem-solving: understanding the problem better, generating more and better ideas, as well as visualizing alternative scenarios for user and stakeholder reactions to new ideas.

Key insights 

A 2022 study by Stevenson et al. tested ChatGPT-3’s creative performance using the Alternative Uses Test. The test presents an everyday object (such as a stapler, a vase, or a soap dispenser) and instructs individuals to think of as many alternative uses for this object as possible (Guilford, 1967). They compared the AI responses to human responses using expert ratings. Their results showed AI responses to be ‘impressive, and in many cases appear human-like.’ However, they concluded that human creativity outperforms that of GPT-3 (Stevenson et al., 2022).

However, just the following year, a new study focused on the creative performance of ChatGPT-4. Guzik et al. tested ChatGPT-4 for creativity using the Torrance Test for Creative Thinking (Guzik et al., 2023). The test requires participants to ask questions, guess causes and consequences, invent product improvements, come up with unusual uses, and imagine an improbable situation (Cramond, 1994). The AI responses were blind-rated and compared to a human control group by professional raters who did not know which were produced by AI. The results were impressive: ChatGPT-4 ranked in the top 1% for originality and fluency, and in the top 7% for flexibility.

Even more interesting, there seems to be a symbiotic relationship between AI and the human user. The user can educate, facilitate, and enhance AI to ‘become’ more creative during a problem-solving process. It takes creativity to enhance AI creativity.
As such, we may see a future where highly creative humans now have access to an enormous creative AI resource, while less creative humans have access to a less creative AI resource.

We do not know how creative the next upgrade of AI will be; however, it is likely that it will be even more powerful in terms of creative thinking. Some of the founding ideas for AI actually included creativity and originality as key features (McCarthy et al., 1955). However, even now, we may receive a more novel response from AI than we would get from our colleagues during a brainstorming session.

Who will you involve to solve your complex problems: ChatGPT-4 or your human colleagues?

Perhaps the question is better phrased: How can you involve AI in your team’s complex problem-solving?

Complex problem solving with AI

It is possible to categorize the purpose of creative thinking into a series of process steps, with some of the more important steps being problem understanding, idea development, and idea evaluation. 

Problem understanding

Identifying everyday problems during work tasks is easy: the machine has stopped working, the team has numerous conflicts, the storage room is too small, etc. Some user-centric problems also reveal themselves quite easily: written complaints from users, customers choosing another supplier, users incorrectly using the product, etc. Quick and efficient solutions are needed to maintain the organization’s functionality.

Sometimes, we aim for more than just ‘functioning.’ We aspire to innovative solutions for a particular problem. Or perhaps a recurring problem demands a new type of solution – one that we might have never considered before.

When dealing with everyday (complicated) problems, we use data and analysis to better comprehend the underlying issue. However, for complex problems, this approach rarely identifies the ‘root cause.’ Complex problems often necessitate entirely new perspectives before we can solve them properly. New perspectives lead to new problem understandings, which, in turn, enable the production of new solutions.

AI can aid in providing new perspectives on the problems we are trying to understand. We can ask it to describe all the problems related to a specific situation, interaction, object, or function. Alternatively, we can provide AI with some problem perspectives and ask it to generate more. For instance, ‘We have identified the following problems related to a potato peeler: difficulties peeling smaller potatoes, challenging cleaning process of the peeler, peel scattering across the kitchen table… Can you please suggest more potential problems related to the peeler?’

We can also task AI with ‘developing 100 alternative problem definitions for the following problem…’ AI can help challenge our current standards—norms, traditions, and ways of working that we overlook. These blind spots can be described to AI in terms of our current work process related to a particular problem, asking it to ‘help challenge these work processes.’

Using AI to help define vague problems into more well-defined sub-problems might be the most potent approach to gain new perspectives on a problem. Well-defined problems for a better hotel experience could include ‘new ways to hand over room keys, innovative departure processes for hotel guests, a reimagined room service process…’ Such well-defined problems naturally lead to more concrete ideas.

Idea development

The better the ideas we develop, the better solutions we create. Therefore, investing time and effort in generating new ideas for the problems we are solving is crucial. We might already have a few ideas, and one of them might even solve the problem. If the problem is less critical, this viable solution might suffice. However, for important problems, we need multiple good ideas to choose from. Relying on a single viable idea provides the weakest possible foundation for decision-making. Generating more ideas builds a stronger decision-making foundation.

AI can assist us in generating more ideas, including novel ones. It can help generate ‘why haven’t we thought about that before’ type of ideas. Therefore, inviting AI into brainstorming sessions or innovative workshops can be valuable. We need to find meaningful ways to integrate AI into complex problem-solving.

We can request AI to ‘produce some novel, unusual, and unique ideas’ for our problem. Alternatively, we can share our existing ideas for the problem and ask AI to ‘generate alternative ideas that solve the same problem.’

If AI responses resemble our existing ideas too closely, we can clarify that we are seeking innovative breakthroughs and request AI to ‘propose only original ideas that 99% of people would never consider.’

Seeking a genuine creative breakthrough might require training AI to advance in creative thinking. Direct AI through a series of creative methods or techniques, such as employing idea combinations. For instance, if brainstorming new ideas for passing an opponent player in soccer, instruct AI to draw inspiration from Tango dance steps. For optimizing supermarket checkouts, suggest AI gather insights from airport security. If reimagining how to try on clothes in a clothing store, prompt AI to draw inspiration from a furniture store showroom. Then instruct it to ‘find other sources of inspiration and generate novel ideas for the problem.’

Idea evaluation

Novel ideas are vulnerable as discussions tend to favour old ideas. It’s challenging to find strong arguments for novel ideas, whereas it’s easier to generate solid arguments for familiar (old) ideas. Hence, new ideas often get overshadowed by the myriad of compelling arguments for old ideas. Evaluating ideas requires a visionary mindset that seeks potential in ideas offering new directions of thinking. Sometimes, new ideas leverage logic from different professions or industries, making it harder to identify their potential.

AI can help envision potential positive consequences if an idea were implemented. It can visualize the idea’s details, operational aspects, implementation process, and user and stakeholder reactions.

We can ask AI to create user scenarios for a novel idea, such as ‘please imagine how serving coffee in tin cans might be like for 10 different types of cafe customer personas.’ We might ask it to imagine ‘how employees in a manufacturing plant might react if we replaced all internal communication, including e-mail, with voice messages.’

We could also request AI to ‘state the pros and cons for the following idea…’ or compile a list of potential drawbacks and benefits if the idea were implemented. However, providing criteria may yield more detailed responses. For instance, asking AI to ‘use novelty, feasibility, and desirability criteria to evaluate the following four ideas against each other.’

Evaluating new ideas necessitates an understanding of the context. AI is capable of providing numerous responses. However, providing background information, detailing priorities, available resources, and constraints can ensure more relevant responses for you and your organization. Be cautious about providing only the fundamental constraints and priorities, as some constraints might be purely cognitive and subject to change when presented with the right idea potentials.

Implications for Businesses 

Innovative organizations put considerable effort into ensuring a continuous flow of new and valuable ideas. They train their employees and leaders in innovative skills, implement idea systems to collect, evaluate, and select ideas, establish support and reward structures for creative efforts, and leverage creative process methods and techniques.

Organizations fostering a culture of team creativity might eventually gain new competitive advantages. Now is the time to contemplate how AI creativity might integrate into this innovation framework. It could be an opportunity to leap forward in innovation, even for organizations that have not yet developed an advanced innovative structure.

Conclusion

The paper explores how creativity drives innovation in organizations and the evolving role of AI, like ChatGPT-4, in aiding creative problem-solving. It emphasizes AI’s ability to generate original ideas and suggests integrating AI into problem understanding, idea generation, and evaluation stages. Businesses fostering team creativity can gain a competitive edge by creatively incorporating AI. It prompts consideration for leveraging both human and AI creativity for maximum innovation potential in problem-solving.

REFERENCES