Ideation and Experimentation : AI and Design Thinking tools for automation

Summary?

In this article I explore the blend of AI and Design Thinking and the potential of applying generative AI. I try out the collaborative landscape within Replit, OpenAI API and building things in the browser to understand how it could help reshape our ideation methods – plus the balance between technological advancement and preserving the human touch in the creative problem-solving approach.

Rediscovering Curiosity – Design Thinking and AI

During my period of redundancy prior to joining Magnetic , I engaged in an exercise to keep my mind active and a much needed break from LinkedIn. I explored the intersection of Design Thinking and AI, simply aiming to understand how these two areas could work together.?The focus was on surfacing how AI could benefit the Design Thinking process, improving and learning the approach to innovation.?

I started my journey with a fundamental query: how could AI be integrated seamlessly into the design approach, maintaining ethical and fair standards? My focus was on using technology responsibly, aiming to enhance our existing methods instead of displacing them.?

The goal was to find harmony between the creativity and AI innovation, ensuring a mutually beneficial coexistence.

Building on this exploration, I turned my attention to our valued "How Might We" (HMW) statements, integral to our design thinking approach. These statements have always been a catalyst for innovative thinking and problem solving. It led me to consider the potential of automating their creation. The question that followed was: Could we develop a tool capable of crafting these statements from specific inputs, thereby amplifying our problem-solving abilities? I probably could have also framed this as a HMW Statement.?

Simple Tools for Complex Challenges: The DIY Approach

During my role as the strategic design lead at Common Good, I led a couple of projects that reshaped a number of pharmacy services. One engagement included partnering with a high street pharmacy to re-imagine the repeat prescription process. The primary objectives were to expedite prescription collection, give time back to pharmacists, and minimise costs associated with uncollected prescription bags and returning medication and breaking down the bags.

To tackle this challenge, I opted for a straightforward approach. Recognising the potential of a browser-based solution, I revisited my basic knowledge of HTML and JavaScript. Drawing from my past experiences in exploring AI applications at a basic level, I identified Replit as a user-friendly platform, ideal for someone like me who wasn't a seasoned developer.

Seeing the OpenAI API as a valuable tool, I focused on its practical application in this simple exercise. With determination and the assistance of resources like ChatGPT to help me code some things that I was unfamiliar with, I managed to get something going in the browser.?

Crafting a minimalist, user-friendly form using HTML and JavaScript, I established a direct link to the OpenAI API. While security concerns lingered around exposing the API key given the project was public, my focus remained on the fundamental proof of concept, emphasising functionality over complexity for this first exploration.

Balancing Functionality and Perfection: Refining the Prototype

The form accepted two inputs: a persona or behavioural archetype and a design challenge or problem statement, our canvas for ideation.

I retained the DIY approach, I jumped into the mechanics of the OpenAI API within the confines of Replit. What got me curious was how seamlessly it functioned, providing responses directly within the browser. Yet, the initial output was far from perfect. The responses appeared jumbled and somewhat challenging to decipher, prompting me to refine my methods, tweaking my interaction with the API for clarity. I turned back to ChatGPT for this and learned more about the API and how to use it in a way that would return what I needed.?

I recognise that as a designer, sometimes, it's okay to do just enough and move forward, especially in the spirit of DIY and quick problem-solving. Take the ordered list of the response with 20 HMW statements, for instance. It worked well, but I found myself tweaking the layout, a detail that might not have needed such attention on reflection.?

I considered incorporating additional inputs such as an organisation's design principles or values. Recognising the humanistic approach these principles could add, I thought of a more holistic approach, allowing us to align AI-generated ideas with the core values and philosophies of the said organisation.

The Power of Collaborative tools –?Ideation with Replit, OpenAI API, and the Browser

In one evening, I managed to piece together a functional prototype. Although the output required some fine-tuning, I successfully proved that the OpenAI API could serve as a valuable tool in our ideation process. Specifically, it proved highly effective in generating a high volume of HMW statements, even though some formatting adjustments were needed.

Despite the need for refinement, these responses would hold significant value for our experimentation and ideation efforts. The blend of Replit, ChatGPT, and the OpenAI API proved to be a surprisingly powerful combination. The learning curve, much to my relief, wasn't as steep as I had anticipated and I’m keen to learn more.?

Two weeks later, I came across Board of Innovation's AI toolkit, which included a well executed HMW statement generator. This confirmed that others in the industry were exploring similar paths, validating the potential of AI-assisted ideation. It was a reassuring reminder that our journey, while mine was individual, could be part of a collective movement toward innovative solutions.

Key Takeaways and Reflections:

Reflecting on this journey, a few key takeaways emerge:

Simplicity Over Perfection: This experience reinforced the power of simplicity. While I spent time refining details, the essence was in the straightforward approach. Sometimes, it's essential to embrace imperfection and focus on functionality, especially in the realm of rapid experimentation.

Collaborative Learning: Leveraging resources like ChatGPT opened doors to collaborative learning. Seeking guidance and support when needed, whether from AI tools or the broader community, enhances the learning curve significantly.

Human Touch in Technology: Integrating behavioural archetypes and organisational principles highlighted the importance of the human touch in AI-driven ideation. By combining artificial intelligence with human values, the creative outputs become more meaningful and aligned with organisational goals.

Questions for Further Exploration:

Where Else Can AI Assist in Ideation and Experimentation? Considering the success with HMW statements, what other aspects of the ideation process could benefit from AI assistance? Could AI enhance brainstorming sessions or aid in user research analysis?

Exploring Ethical AI: How can ethical considerations be further integrated into AI-driven ideation? Ensuring that AI tools align with ethical principles and values is essential. HMW strike a balance between innovation and ethics in this domain?

Scaling Collaboration: How can collaborative efforts, like those between designers and AI tools, be scaled across diverse teams and projects? Exploring methods to democratise AI usage in creative processes could lead to more innovative outcomes.

Embracing the journey of experimentation and learning, I’d love to hear what you think could be answered in the comments. There are definitely more uncharted territories where human creativity and artificial intelligence converge could help? shape the future of ideation and innovation.

Thanks for reading and let’s make something great together.?