User research – the systematic investigation of user needs – is now an essential part of good UX and product development. Interviews, tests, and observations help to understand real requirements, identify priorities, and develop relevant solutions. But this process takes time and money – two resources that are in short supply in many projects.
This is where a new trend comes in: AI personas and so-called synthetic users. They promise to make target groups quickly accessible and deliver research results in minutes instead of weeks. But does this approach deliver what it promises?
AI personas are created using generative models such as ChatGPT, Gemini, or specialized platforms. When you enter a rough description of your target audience, the AI generates typical parameters such as age, income, habits, values, or social media behavior. This results in profiles that serve as representatives of a target audience. With the right prompt wording, these personas can also be fine-tuned and condensed.
Synthetic users go one step further. They generate not only profiles, but also entire interviews. When you enter a research question, these virtual users respond like real interview partners – including biographical details and pre-formulated statements.
At first glance, such results are reminiscent of classic research artifacts. But unlike real personas or interviews, they lack the human context of experience and observation.
For B2B companies with complex products and long decision-making cycles, this quick start can help outline initial requirements and prepare workshops.
However, research by the Nielsen Norman Group (NN/g) – an internationally recognized consulting institute for usability and user experience – shows how distorted the results of synthetic users can be.
Synthetic users are no substitute for real interviews – but when used correctly, they can provide valuable services. We have put together a few potential scenarios for you:
A real-world example: A leading US insurer uses Dynatrace Synthetic Monitoring to continuously test critical applications. Simulated logins and transactions check availability and performance – and report problems before real users are affected. In addition, the company uses synthetic tests on its competitors' login pages to compare its own performance in the market environment.
When resources for real interviews are lacking, synthetic research can be a helpful starting point. However, there is a risk that inaccurate results will be accepted without verification – or that teams will become accustomed to quick availability and permanently dispense with real research. Exclusively using synthetic research can therefore prove to be a mistake in the long term.
AI personas and synthetic users are tools with potential, but they must not be misunderstood. They are suitable for initiating projects more quickly, visualizing hypotheses, and activating teams. However, they cannot replace the depth, empathy, and authenticity that only direct contact with real people can provide.
1. Use synthetic users for getting started and forming hypotheses, but always validate results with real research.
2. Use them where scaling and monitoring are required – for example, for load testing or market benchmarking (i.e., systematic performance comparison with competitors).
3. Avoid using synthetic users as the sole basis for decision-making in critical projects – they are a supplement, not a substitute.
Would you like to know how your company can use AI to complement user research in a meaningful way? Uhura Digital supports you with strategy, research expertise, and the development of digital experiences that really reach users. Contact us for a no-obligation initial consultation.
