What it's about: Companies must understand user needs in order to remain competitive. yosona enables designers to interact with AI-personas to gather perspectives, test ideas, and develop hypotheses, complementing real user interviews.
Companies must continually evolve their offerings to remain competitive in a complex and ever-changing world. It is essential to identify and incorporate user requirements in order to design solutions that address real needs. Human-centered design, for example, provides a range of practices for developing people-focused products and services. For designers, it is crucial to engage with their target audience – typically through interviews, observations, and ultimately user testing. These methods are vital for understanding users’ needs and desires and for developing a design solution that effectively addresses their problems.
Because of limited resources, time constraints, and the limited availability of the target audience, this direct contact is typically restricted to only a few points within the product development process. Especially during the actual development phase, there is usually no direct exchange with users or other stakeholders – that is, all individuals, customers, and organizations that influence the product or service. To stay continuously oriented toward users nonetheless, designers rely on various methods, including personas. These are fictional yet realistic descriptions of (potential) users of a product or service. They help the design team adopt the perspective of their user groups and form a clearer understanding of which solutions would best support them. In doing so, the design team draws on existing knowledge about the target audience as well as on their own experiences – and thus also on their own biases.
Instead of relying solely on their own imagination, designers can consult AI-personas. In this approach, a large language model, such as ChatGPT, is instructed to assume the role of a persona and communicate as a real user would. To do this, the AI draws on a wide range of publicly available information, project-specific data, and the underlying persona profile.
Using AI-personas offers several advantages: It makes it possible to incorporate the perspectives of multiple user groups in a short amount of time. In addition, the likelihood of gaining new insights increases through the AI’s unexpected prompts and responses. Large language models can represent human experiences on a broad scale and help designers develop a deeper understanding of the user perspective. They can also act as a “partner” in idea generation by introducing new ways of thinking. One major benefit is that AI-personas can simulate the viewpoints of different user groups at any time and at nearly any stage of development.
In design projects, the needs and opinions of various stakeholders such as users, businesses, or public authorities often have to be taken into account to ensure that a product or service is viable in the market. To use personas for this purpose, it is helpful to be able to communicate with several AI-personas in one chat and quickly understand which views they share and where they differ. Existing AI systems such as ChatGPT are not designed to efficiently gather the perspectives of multiple AI-personas, meaning different user viewpoints, or to help users keep track of them. Creating personas within these systems is also not ideal.
The application yosona developed in this project addresses these challenges. It allows users to create AI-personas and assign them to different chat groups. This enables them to communicate with several AI-personas at the same time within individual chats. The tool helps users identify similarities and differences in the personas’ responses at a glance by automatically highlighting the topics they mention. In this way, designers can quickly gather and compare different perspectives.
Let us look at the fictional company Wellory Health Tracking as an example. The company offers wristbands that measure the vital data of their wearers and help them improve their personal health in a targeted way.
Wellory wants to continuously improve its app so that the recorded vital data is easy for users to understand and so that they remain motivated over time. This is where yosona comes in. The Wellory team has created several AI-personas in yosona that represent different user types, including busy people with little time, health and fitness enthusiasts, and individuals with specific health needs. In chats, the team gathers feedback from these personas on new ideas or features. Their needs and challenges are also examined in greater depth to develop important assumptions that influence how the app may be used, which can later be validated with real users.
As soon as the designers at Wellory have created initial concept drafts or a sketch, they upload them to yosona. The AI-personas respond to the material according to their profiles, in other words in the way a person with those characteristics would respond. In doing so, they point out potential problems or possible improvements.
To better understand the basis on which the AI model forms a persona’s opinion, users can activate a “transparency mode.” In this mode, the profile information that the system used to generate the response is linked. This allows designers to see whether the persona’s conclusion makes sense and aligns with the established profile. If the design team is working with validated personas, meaning personas based on real user research, it can rely on robust data. If no suitable reference is found, the team knows that this aspect still requires closer investigation. This way, they can immediately see where open questions remain and address them in a targeted way when real users are involved next time. In this way, yosona makes existing knowledge visible exactly where it is needed and helps designers identify relevant gaps in understanding.
With every new question, feature, or design sketch, the personas can be brought back into the conversation, which keeps the team in constant exchange with its target groups. While real interviews and user tests remain valuable, yosona closes the gap between the major project steps. In this way, companies like Wellory can continuously and quickly optimize their products and services without losing sight of the user perspective.
yosona is an application designed specifically for working with AI-personas and enables designers to quickly gather feedback from the perspective of their target audience. The focus is on capturing different viewpoints and making them easy to compare. Designers are supported in identifying the distinct and shared opinions of the personas, which helps them determine the most important aspects for developing a product or service.
The use of AI-personas in human-centered design and similar practices is a new concept, and there is not yet a clear strategy for how they should be applied. For this reason, a set of guidelines has been developed to help designers understand when and how they can use AI-personas in their projects.
Individual project
In cooperation with Fraunhofer IAO
Supervision
Prof. Carmen Hartmann-Menzel
Dr. Matthias Peissner