From skepticism to synergy: My experiences with AI in UX design

You have probably heard the saying, "AI won't replace your job, but people who use AI will". When I first encountered ChatGPT and the transformative potential of generative AI in January, I must admit, my initial reactions were skeptical and even somewhat dismissive. I protected my UX territory by saying things like, "Change our world? It can never truly match the skills of a UX professional, it will only make us lazy, it lacks creativity, and surely, the time we currently commit to our tasks is too important to be ruined with AI…". Yet, despite the validity of some of these doubts, they did not stop me from exploring what has since become a key element in my role as a UX designer and researcher.

Katrine Ly
—UX and Insights Manager

My initial application of ChatGPT for idea generation and concept development

Acknowledging that AI is here to stay, I chose to plunge into generative AI through hands-on experimentation. Obviously, ChatGPT became my vehicle for this initial exploration.

Yet, it wasn't until I used it for a specific project – a complete redesign of an engineering company's website – that I saw its true potentials and limitations. Despite knowing little about the engineering business, I asked ChatGPT for insights into this specific company and its competitive market to establish solid knowledge and kick-start idea generation. ChatGPT delivered extensive information about the engineering business, such as insights on both national and global market standings, a rundown of various segmentations, and a clear explanation of the different types of engineering. All of this information played a critical role in shaping my initial ideas for the overall digital concept and the market positioning strategy.

Later in the process, I leveraged ChatGPT to create personas and user scenarios. With well-considered prompts, it generated a rough draft for four distinct personas representing our key user segments, along with two tailored user scenarios. I validated these personas and scenarios, and they served as my foundation for creating hero task flows (simplified journeys). Ultimately, these personas, scenarios, and hero task flows helped me sketch out and define the website's overall structure, including its five distinct page types, along with the necessary experience drivers and required features.

During concept development, ChatGPT aided in brainstorming labels, call-to-actions, headlines, and copy. It helped clustering various topics into coherent categories for our four user segments. The client didn’t only validate the categories, but was thrilled about the final concept. This marked a satisfying conclusion to my first real application of generative AI.

Utilizing various AI-tools for research and insights management

My experience with ChatGPT piqued my interest in exploring other AI tools, and more specifically those geared towards research and insight management. While ChatGPT is a brilliant tool, it has its limitations and isn't designed to support thorough research processes that require validated insights for informed decision-making.

It is estimated that 19% of the work week is spent finding information and 55% of all data and insights remain unused. (Sources: ESOMAR, McKinsey, CMI Interviews)

I am part of a dedicated team of strategists and digital UX designers, all holding the belief that understanding people, along with the organizational past, present, and future, is essential for informed decision-making, formulating clear strategies, and crafting unique digital solutions. However, this approach relies heavily on having accessible knowledge, data, and insights, which is far from the reality today.

Thankfully, a growing number of innovative companies have recognized this challenge and developed AI-powered tools to transform disorganized data into structured knowledge and insights. For these purposes, we have been privileged to use Stravito, and we also use Custom GPT and Dovetail.

Stravito: We utilize Stravito as our central AI-driven knowledge base for individual projects, providing us instant access to specific information and the possibility to create unique scrapbooks of essential insights. 

Custom GPT: Alongside Stravito, we use Custom GPT for storing knowledge and data for individual projects. Its unique feature is a customized GPT4 chat that responds exclusively based on the data uploaded to the platform. This means Custom GPT – as opposed to ChatGPT – ensures accurate chat responses with precise source references.

Dovetail: Finally, we utilize Dovetail for analysis and the creation of new insights. Dovetail provides speedy and surprisingly accurate transcriptions of interviews, helps to create summaries of large reports, and to a certain extent, highlights key insights for analysis. 

While I cherish all of these above-mentioned AI-powered tools, my initial skepticism remains intact when it comes to the analysis and generation of new insights. I am still convinced – and perhaps a bit stubborn, that these steps require RI (real intelligence) human judgment, perception and creativity. More importantly, I believe that they are essential for us as UX designers in our process of turning knowledge and insights into initial strategies, conceptual thoughts, and concrete ideas. 

Continuing my looping journey from skepticism to synergy

Despite the undeniable benefits of AI, my journey has also emphasized the irreplaceable value of real intelligence of human judgment, perception, and creativity – particularly when transforming knowledge into insights, strategies, concepts, and ideas. 

AI may be a powerful tool, but it is just that - a tool. The expertise, creativity, and empathy of UX professionals truly bring our digital solutions to life, ensuring unique user experiences and innovative solutions. That said, I stay curious, and continue to explore new methods, tools and features within the evolving landscape of AI, always ready to be surprised. 

Katrine Ly
UX and Insights Manager