back

Generative AI for UX – News from the design front

September 13, 2023

We have already presented many ways in which Generative AI will turn the digital business world upside down or where this is already happening. So far, we have focused on programming, after all, many of the features offered by tools like ChatGPT out of the box or as plugins are optimized specifically for these use cases. However, this is not the only field in which Generative AI is expected to make a change. In UX, tools already exist to give designers and researchers an edge in many areas of the design process to work more efficiently and achieve better results.

Brainstorming & inspiration

With the entire Internet at its fingertips, Generative AI tools are a great option to help out when a bit of inspiration is needed. Whether a vague idea of the final product already exists, or a rough collection of references should be created first, artificial intelligence can deliver a multitude of ideas in a matter of seconds, which can then be evaluated and curated. Tools such as Midjourney or Adobe Firefly are particularly useful for this. The fact that details such as icons or texts are often difficult or even impossible to recognize is not important for the time being. This step is primarily about gathering impressions of a possible overall appearance. 

Content generation

Anyone who designs products is familiar with Lorem Ipsum, the completely meaningless placeholder text that serves to direct the focus away from the content and onto the overall appearance. However, especially when it comes to presenting high-fidelity designs to customers, it can be helpful to rely on real content to paint as realistic a picture as possible. However, this content is often not yet available. AI tools can help without distracting designers too much from their core business. Among the use cases are product descriptions or longer paragraphs, created for instance with the help of the Figma plugins Relume or Magician. Images too can be created reliably by using AI: Tools such as Firefly are particularly suitable. Plugins such as Summon.AI or the mentioned Magician make it redundant to even leave the design tool. If necessary, this content can then be adapted and optimized by real people. Especially in the area of text generation, tools developed specifically for this purpose are now springing up like mushrooms - but useful results can also be achieved quickly with ChatGPT on-board tools.

Wireframes and UI Design 

Especially when it comes to generating wireframes, which are often not about detail but about implementing a rough concept, Generative AI promises to be a helpful companion. Plugins such as WireGen can generate wireframes in a matter of seconds. The rapid iteration through different design approaches can be AI-supported in the future. Already, plugins exist for relevant design tools that promise to do just that. Prominent representatives are, for example, the design tool Uizard or the no-code website builder Framer after the integration of a corresponding AI feature. Figma's recently announced acquisition of Diagram, which is at the forefront of AI design tool development, shows that the big players in the business have also recognized the potential and are charting a course toward AI.

Research & Testing

Research and testing of designs are integral steps in the design process. However, extensive usability testing often consumes resources that are not always available. AI can be particularly helpful where it is necessary to confirm or disprove fundamental assumptions about user behavior. Although the field of UX research is not expected to be fully automated in the near future, many tools already exist that offer automated executions of a wide range of methods. Many of them rely on OpenAI technology for their services, for instance, to answer surveys or to investigate copy text in A/B tests. AI can also be used in the evaluation phases, for example, to create transcripts of user interviews. It remains to be seen to what extent AI tools will be able to establish themselves in this field.

Personalized Experiences

Tailoring digital products to the needs of users is one of the main goals of UX design. Especially when it comes to analyzing large amounts of data, machines already perform better than humans. This opens up new possibilities for personalization, for example when it comes to suggesting content. The streaming giants Netflix and Spotify have already recognized this and are using generative AI to make better suggestions for their users.
However, it's not just content that will adapt to the individual user in the future, but also the interface between man and machine itself. So-called adaptive interfaces are nothing new per se: To meet the requirements of different markets, designers have always had to make adjustments to digital products - a simple example of this is the reading direction from right to left in Arabic languages, which have to be displayed differently in digital interfaces than languages that are read from left to right. Artificial intelligence, however, offers the possibility of analyzing user behavior more precisely over long periods of time and adapting the interface to this behavior. Some time ago, Siemens implemented this form of adaptive UI in their CAD program NX in order to better meet individual workflows - we can assume that more examples will follow.

Conclusion

Generative AI is already turning many areas of the product development cycle upside down - user experience design is no exception. We got to know the fields where work can be done AI-supported already. These tools will evolve, and new ones will surface.
However, not only the process but also the interfaces themselves will change. With the rapid spread of AI applications, there is a new need for suitable interaction options to achieve desired results. These interaction paths also need to be designed. In the future, UX designers will not only use AI tools for their work but will also have the opportunity to influence the development of these tools in favor of the needs of the users, thus fulfilling their mission as "user advocates".

Read more...