How AI Will Change UX Research
The time is now to level up your AI skills
In the age of AI, everyone has their predictions of how industries and job functions will change. I’d like to give my two cents on how AI will change UX research, or more broadly, how it will change how product teams understand customer data and react to it.
On one end of the spectrum, there is the doom and gloom scenario. For many professions, AI is a scary prospect. It threatens to upend livelihoods and force workers to learn new skills from scratch.
In this scenario, AI takes off and can replace the operations and execution of UX research. Companies will only need one research expert on staff to give a prompt to an AI. The AI will search internal documents and external sources to find the answer to the question.
AI will leverage facial data as users navigate through an app to predict future behavior without ever asking a user for their opinion. AI will identify and fill gaps in understanding based on its own learnings, conduct its own interviews and surveys, and fill in with synthetic data where feasible.
It will reduce the decision-making process when it comes to concept testing multiple UI designs. Instead of running research or A/B testing, it will analyze designs and select which one will perform better based on millions of data points it is already trained on.
All the researcher will need to do is ask the right questions. The AI will do the rest. While most of these use cases sound more sci-fi than reality, some may come to fruition eventually. Still, it is likely years down the road.
In reality, AI will likely displace a portion of the UX research workforce, but not all of it. It will shorten workflows and make researchers more efficient. Researchers will be able to have more impact on decision-making, provide more data points for product teams, and drive better outcomes for the business and users.
My bold prediction is that qualitative research will become more important in a world filled with AI. Product teams will be itching for more emotional intelligence, user insight-driven innovation, and ethical oversight in product development, which will be more difficult to replicate with existing datasets that LLMs and AI models are built on.
I’ll get into why these three things will still be important, but first, how will AI actually change the operations of data gathering and data analysis for UX researchers?
Replacing Mundane and Operationally Intensive Tasks
Data Gathering
AI is already able to help with data gathering in some ways. The most obvious way is to prompt an LLM to help with creating an interview guide and survey design. LLMs can give researchers a great starting point—something like 75% of the way there in most cases—to reduce the amount of time and brain power needed. But the prompt is important. Researchers need to have the key research questions in mind to get relevant content back that isn’t just 90% of the same questions redesigned for a specific product.
Part of the data-gathering process is also ensuring you are gathering data from the right user segments. AI can pull customer lists from company databases. Soon, there will be a way to prompt, “I want a list of users with A characteristics, B behaviors over the last month, and C demographics.” Instead of having to write SQL code, the AI will be able to easily pull the list based on plain English instructions.
When it comes to recruiting participants for a study, AI will be able to craft emails or in-app messaging and adjust compensation amounts based on real-time feedback and response rates. Once participants are found, the AI will be able to schedule participants for moderated qualitative studies based on who most closely aligns with the study requirements, without the intervention of a human.
Eventually, once humans are more familiar with naturally interacting with AI, AI avatars will be able to conduct interviews directly with participants. In this future, the most important piece of qualitative research will be the creation of the interview guide.
Data Analysis
AI will be a powerful tool for data analysis. Researchers will use AI to run power analyses and get recommendations on which statistical methods to use based on the desired insights. While it can do this now, it can get so much better.
In analyzing qualitative data, AI will be able to find patterns and put probabilities on when someone is telling the whole truth versus a half-truth.
For quant datasets, AI will not only be able to respond to prompts to perform different statistical analyses and models, but it will also be able to identify trends and unique insights in the data without specific prompts to do so.
What AI Won’t Replace (At Least in the Short-Term)
Emotional Intelligence
AI doesn’t currently have emotional intelligence. Identifying different types of emotions, being empathetic, and building rapport are things AI can’t currently do. Skilled qualitative researchers still need highly refined soft skills and processing capabilities to translate users’ emotions into beneficial insights for product teams.
In a world of AI, researchers who are educated and trained in behavioral and social sciences will become more important than researchers trained in statistical modeling. Understanding and knowing how and when to use statistical methods will be important, but reasoning and math are far easier for AI to replicate than generating qualitative insights.
User Insight-Driven Innovation
AI bots and LLMs are built from existing data. We will still need human minds to take what we learn from research and existing data and find ways to innovate for future products. AI doesn’t know what it doesn’t know, but humans can use their imagination and ideas to bring new concepts to life that AI can’t.
Ethical Oversight
AI will mainly be leveraged to optimize for shareholder value and product success. It won't be incentivized to think of the user first and err on the side of caution when ethical or moral concerns arise. Humans can step in and use common sense and a user perspective to drive product changes and innovations while still implementing ethical oversight when an AI’s job may not do so.
Key Takeaways
If you are a UX Researcher or working on the voice of the user in your organization, don’t fret. I don’t think you will be losing your job over night because of AI. First, AI will make you more efficient. Then, it will make your skillset more valued. My advice would be to embrace these new technologies. Play around with them. Find ways to implement AI into your workflows. Become comfortable with it and find ways to gain an edge over the competition.
Then, up-level in the areas of research you feel you may be lacking. If you are a purely quantitative researcher, take on qualitative projects. Learn how to conduct an interview. Understand the nuances of analyzing qualitative data. If you are a qualitative researcher, learn the language of statistics. As AI becomes more prevalent, more researchers will likely be required to use mixed methods and be more efficient. It’s a great time to be a researcher and dive into customer data. Embrace the AI revolution.