Lessons learned from leading a design team during the Gen AI explosion (Part Two)

Llara Geddes is a Solutions Director at AND Digital. She has a background in UX and generating actionable insight to inform strategy and solutions for clients.

Part One covered AI and Research. In part two of Llara's blog she brings us continued insights and expert tips from leading a design team through a generative AI explosion.

Llara continues...

"Revisiting insight

One issue we come up against in research is how to build an effective repository. I’m seeing more and more teams using tools like Notebook LM to upload research transcripts, so that teams (with context - Gate 3) can go back and question and chat about insights and research findings, to great effect. Utilising Notebook LM in this way also allows teams to output video and ‘podcast’ summaries of findings. While I don’t advocate for using this as your sole output, they can be very effective in engaging stakeholders who want a high level view, or who will never read the deck or engage at a deeper level.

Effective use of AI in Research

When it comes to broader market/demographic research, there are some good use cases for AI. Tools like Gemini’s Deep Research can save time and accelerate your understanding of sectors and demographics. Take care to apply Gate 2: fact check sources and claims to make sure they stand up to scrutiny.

A lower risk approach to AI in research is in using it to remove blank page paralysis when creating research plans and writing scripts. By using well crafted prompts (Persona, Task, Context, and Output Specification) (Precision Stack or Four-Layer Prompting Framework) and even feeding previous or templated plans, you can get a head start. This work should only ever be treated as a starting point, with you, the human reviewing and editing to meet your needs. Do not rely solely on AI output.

AI and Design & Prototyping

Designers: AI is not taking your job. Not right now. The crux of great design is that it solves problems. It takes clearly understood needs and crafts solutions that make information clearer, life easier and experiences more delightful.

Design & Prototyping

When people think about AI and design, they jump to tools like Figma Make, Lovable, Bolt, V0 and so on: tools that ‘design’ interfaces and create prototypes. These tools are great for expediting some processes, and with a brilliantly crafted prompt, or through trial and error, you can get some decent design from them. Integrating your design system further enhances the visual results you can achieve with them. However, expecting to just hop into an AI prototyping tool and get all your work done is misguided. The value designers bring is in their thinking and deep understanding of the user problem.

Where these tools often excel is in expediting the process to get to prototypes to test flows and validate concepts. The speed with which you can pop something together that’s ‘good enough’ to run testing is impressive. And this speed allows us to test multiple flows and concepts. Previously, we’d design, prototype, test. Iterate, re-design, prototype, test. Now we can get a prototype together in a fraction of the time. We can have multiple prototypes to test and compare in a single session. And we can even craft sessions to allow for iterating on prototypes on the fly, so that the improvements we identify can be validated in a single testing session.

Ideation

As well as allowing for speed in prototyping, utilising AI tools can accelerate and broaden ideation. Prompting your favourite tool to provide 10 different ideas to solve a particular problem both gives you volume and removes blank page paralysis. Again, it is critical to use your skills to determine the potential validity of ideas. Do they actually propose viable solutions to the problem?

A similar process can be used to generate hypotheses. Don’t forget to include a Persona, Task, Context, and Output Specification in your prompt for best results. And, again, review the results with a critical eye - you might pick and choose. You might refine your prompt.

Conclusion

There are lots of ways we can use AI in the research and design process - this list certainly isn’t exhaustive. We get best results when using AI as an assistant that can expedite manual work, support idea generation and help with analysis and revisiting insight. It is critical that you keep humans in the loop, fact checking, validating and providing context.

AI cannot replicate human taste, context or understanding. Your thinking is still vital, as is your design eye."

From more from LLara, check out her Medium