Most people who start using AI carry the same bad habit into it. They jump to a solution before they understand the problem. AI makes that habit worse, not better, because it can hand you a plausible “solution” in seconds. That was the opening provocation of our recent webinar, “Your Second Brain: AI Tools Every UI/UX Designer Needs to Work Smarter.”
The session set out to correct a common misread of AI in design. The tools are not the second brain. The second brain is a trained way of thinking working together with tools that know when to step in. The webinar built that argument across three parts: how designers think, which tools fit which stage, and how the two combine across a full project.
How Designers Think
The first part made a deliberate choice to talk about design thinking before touching a single tool. The reason is simple. AI only produces useful work for someone who already has a strong design framework. Without that framework, it produces generic output faster.
Design thinking is not linear. Designers move back and forth between stages, iterating several times before a problem is properly understood. The session anchored this in the discipline that separates junior from senior designers: problem framing.
Three ideas carried that section:
- Symptom vs root cause. A user complaint is usually a surface symptom, not the actual problem underneath.
- Reframing as a skill. Senior designers redefine the problem instead of accepting a brief at face value.
- The Double Diamond. Two diverge-converge cycles, one to understand the problem, one to find the solution.
The webinar walked through the real shape of this work: a narrow or misdirected brief comes in, research reveals what is actually happening, the problem gets redefined, and that reframing changes the final design decisions.
All of this happens inside the designer’s head. Research, synthesis, reframing. The catch is that the human brain has limits. Working memory is small, bias creeps in, and processing large volumes of data is slow. That gap is where a second brain enters. Not to replace the thinking, but to extend its capacity.
Which AI Tool for Which Stage
The most practical part of the session dealt with tools. The point was not to list whatever AI is trending. It was to map specific tools to specific stages of the process.
One principle framed everything that followed. AI tools work like a very fast research and production assistant, but they have no taste, no business context, and no accountability for the decision. The designer supplies all three. Kattu called this creative ownership.
With that boundary set, the session mapped tools to stages:
Research and synthesis (Discovery/Empathize)
AI summarizes, clusters, and finds patterns across raw data like interviews, surveys, and reviews. Useful for automatic interview transcription and summaries, affinity mapping, and sentiment analysis across large volumes of feedback. Tools: Claude, ChatGPT.
Ideation (Ideate)
AI multiplies idea variations quickly and breaks fixation on a single solution. It can act as a brainstorming partner, help draft How Might We statements from research insight, and generate microcopy alternatives. Tools: Claude, ChatGPT, FigJam.
Wireframing (Ideate)
AI speeds up the first visual draft, generating layouts from a text description or reference and producing UI component variants. Tools: Figma Make, Google Stitch.
The do’s and don’ts came back to creative ownership. Use AI to increase volume, not to decide final quality. Always validate AI output against real data or real users. Push repetitive and administrative work onto AI so designer time goes to strategic decisions.
And the hard lines: don’t paste AI output straight into a design decision without filtering it through business context, don’t let AI stand in for real user research, and don’t ship raw AI generation as final UI without adjusting it to the brand and design system.
The Live Demo: Designing ShieldView
The abstract argument got concrete with a working case study: ShieldView, a SIEM alert and case management dashboard for SOC analysts. It’s a deliberately hard example, because a security operations brief is dense and technical, far from a designer’s native language.
The demo moved through four stages:
- AI translated the technical SOC brief into designer language.
- It pulled themes, the root problem, and a How Might We statement out of analyst quotes.
- It generated five to six feature concepts plus enterprise UI pattern references.
- It structured three core screens: Dashboard, Alerts, and Case Detail.
The recap made the important point. At every stage, a human still made the call. Which themes to merge, which concept to pick, which element to bring forward. That is the part AI does not replace.
The Full Design Process, End to End
The final part extended ShieldView from a demo into a full story, Discovery through Handoff. Reusing the same case created the intended effect: the audience already knew the context, so seeing the complete picture landed harder.
The session laid out the division of labor stage by stage. A useful detail: today’s live AI demo covered Discovery through Wireframe. From Hi-Fi Prototype onward, the work is recommended to stay manual, led by the designer.
| Stage | Human | AI | Critical decision |
| Discovery | Interview and read user context | Summarise transcripts, clear insight | Real insight vs noise |
| Define | Write the problem statement | Draft HMV statements | Problem priority |
| Ideate | Choose direction by strategy | Generate concept variations | Pick 1 to 2 directions |
| Wireframe | Structure and navigation flow | Generate initial layout | Information hierarchy |
| Hi-Fi Prototype | Adjust brand and interaction | Generate styling and UI components | Ready for user testing |
| Test | Design scenarios, read reactions | Summarise results into insight | Continue or pivot |
| Handoff | Technical discussion and negotiation | Auto-generate documentation | No detail lost |
Each stage connects to a business or user outcome, and that connection is what separated this session from a tools tutorial. Faster research means earlier design decisions and shorter time-to-market. Wider ideation raises the odds of finding a better solution than a competitor. Cleaner documentation reduces miscommunication with developers, which means fewer post-launch revisions.
The framing that tied it together: AI does not shorten a good design process. It shortens the repetitive parts, so the time left over can go to the work that genuinely needs a human. Empathy, judgment, and responsibility for the decision.
Key Takeaways
The real second brain is not the AI tool. It’s the combination of a trained way of thinking and a set of tools that know when to be used, working together consciously at each stage of the process. Or, as the session closed: AI is only as strong as the thinking framework of the person using it.
Zentara’s design and engineering teams build tools like ShieldView for real security operations, where the interface has to hold up under pressure and every decision traces back to a mechanism, not a guess.
If your team is working through where AI fits in your own design or product process, that’s a conversation we’re glad to have. Talk to our team.
