I recently gave a talk on AI for UX designers and graphic designers at a design training centre. This article brings together the points I covered: where AI can help UX designers and graphic designers in their work.
How are UX designers using AI today?
The biggest change I see is in the work that happens before and after research. An interview guide, a first draft of a questionnaire or the structure of a research plan can come together in minutes, leaving the designer more time for real conversations.
The same applies after the research. Processing ten interview transcripts, a pile of survey responses or customer feedback used to take days. AI helps patterns emerge more quickly, while the designer still makes the decisions, because they understand which pieces of feedback point to a genuine problem.
A third major area is rapid prototyping. An idea can now become a clickable prototype in a few minutes, ready to show a client or a user. This makes experimentation less expensive: you can explore more directions before anyone commits to one of them. The design work itself still needs to happen.
Synthetic users: where do they work, and where do you need real people?
One of the most interesting and widely debated topics today is the synthetic participant: an AI-simulated user you can ask questions. In my talk, I used a simple example to explain why we need to be careful with this approach.
Ask seven synthetic participants how many online courses they have completed, and there is a good chance all seven will say they finished every one. Ask seven real people, and perhaps three will say that. The others will explain why they stopped halfway through: they changed jobs, or the content was no longer relevant. The real reasons emerge from the edge cases and exceptions, while the model tends towards averages and a polite yes.
Where can synthetic research be useful?
Exploration and pre-testing. Is a questionnaire understandable? What blind spots might a concept have? Which ideas are worth exploring further? Here, a synthetic participant can be a quick and useful sparring partner.
Comparison, with limitations. Ranking concepts or selecting a direction. AI can help, but the results should be checked with real users.
Where it should never replace people: actual behaviour and decisions. Usability, proportions, recall, emotions and vulnerable groups. These require real people, through observation, interviews and user testing.
If you use synthetic data more extensively, it is worth building in calibration: compare the generated responses with a small real-world sample, and document what each result was measured against, its margin of error and what it can be used for. Nielsen Norman Group’s AI at Work study guide is a useful starting point if you want to explore the subject in more depth.
Services that adapt to the person
There is one idea I think every designer should consider. Until now, people have adapted to services: everyone received the same customer journey, the same amount of information and the same pace. AI opens up the possibility of services adapting to the individual, offering a different pace, level of detail or path to those who need it.
This creates particular opportunities for neuroinclusive design. Someone who finds too much information overwhelming can be offered a simpler path, while someone who understands things through detail can explore further. This requires designers who understand people and know what is worth personalising.
How is AI changing graphic designers’ work?
Image generation is the most visible change in graphic design, but its impact extends much further. It changes how we explore ideas, present a concept to a client, and reach a version we can develop further.
The role is shifting increasingly towards art direction. The graphic designer defines the visual direction, selects the appropriate tools, evaluates variations and brings the result together. Composition, typography and brand knowledge remain just as necessary as before. At the same time, the designer can explore many more possibilities.
Finding promising ideas sooner
In website design, a great deal of time used to go into carefully arranging boxes, building the page structure and drawing the first presentable version. By the time the client saw anything, the designer had already invested substantial work in that one direction.
With the right AI tool, you can now create complete website concepts and even functional prototypes. You can explore the same content with different layouts, visual priorities and imagery. It becomes clear much sooner which direction might work and which is worth letting go.
You can then take the selected version into your familiar design environment. Some workflows turn an interface displayed in a browser into editable Figma layers, allowing you to continue refining the details there. Figma already supports bringing code-based interfaces into an editable design environment.
Sketching by hand and arranging boxes can still have a place if they help you think. But there are now many more routes to that first testable version. The time saved can go into comparing alternatives and developing the strongest direction.
AI-generated work can be edited further
Another important change in graphic design is that AI-generated output is increasingly a starting point you can continue working on. You can generate vector graphics, export them as SVG files, and continue editing them in Illustrator, for example. Adobe Firefly offers this kind of editable vector output.
This allows a designer to refine the proportions of a promising illustration or graphic element, adjust its colours and bring it into line with the brand’s visual system. Transferring an editable website design and exporting SVG graphics are different workflows; what carries over, and how well, depends on the tool.
Professional refinement remains essential. Typography, consistency, performance at different sizes and the requirements of the final use all need checking. There are many important decisions between an impressive first version and a design that is ready to hand over.
Visual direction and selection become more valuable
Here is an example from my own work. Visitors to my homepage are greeted by a bubble animation. I generated the “skins” of the bubbles in Midjourney, placed my own photographs inside them, and we built the movement in code. AI took a great deal of production work off my hands, while the idea, the selection and the judgement about what looked aesthetically right remained mine. Many versions went in the bin because a particle-based liquid effect, for example, looked rather slimy. A designer brings that judgement.
I see the same thing with brands. When a graphic designer builds a brand’s visual system and supports AI use with suitable references, templates and workflows, the rest of the team can work more consistently. The designer’s role also includes defining which assets fit the brand and where professional review is needed.
When a client arrives with an AI-written brief
AI also introduces new tensions into the relationship between clients and graphic designers. A client may arrive with a complete brief, mood board or website design created with an AI tool. This can be a useful starting point: it makes it easier to show what they have in mind and what they like.
The difficulty begins when the client treats it as the final solution and expects the designer simply to execute it. A detailed brief can easily create an impression of professional rigour, even when it leaves out important information about the audience, the business objective or the circumstances in which the design will be used.
The designer then needs to understand the intention behind the proposed solution. What is the client trying to achieve? Why did they choose this direction? Which parts matter most to them, and why? These questions can lead to a collaboration that makes room for both the client’s ideas and the designer’s expertise.
AI does not make anyone an expert
I think this is one of the most important lessons in the whole shift. AI makes it easy to produce something that looks acceptable quickly. Without professional criteria, it is just as easy to stop at a generic result that looks much like everything else.
A graphic designer sees different things in that same result. They notice a weak visual hierarchy, awkward proportions, inconsistent imagery or a design that simply does not fit the brand. They know what to keep, what to regenerate and which parts are best refined by hand.
In the hands of a skilled graphic designer, the same tool can feel like a magic box. Their knowledge, experience and taste help turn those possibilities into something distinctive and valuable.
What stays with the designer
Empathy, good questions, conversations with real people and taste. Understanding the task, choosing the right direction and taking professional responsibility for the final result.
AI speeds up a great deal of work. How much better the outcome becomes depends on what we do with the time it frees up and the possibilities it opens. A skilled designer can explore more directions, experiment more boldly and develop the work they eventually hand over to the client more thoroughly.