Practical Use Cases for AI in UX and Product Discovery

There’s a lot of AI hype in tech these days, and it seems like teams fall in one of two extremes — those who avoid it entirely, and those who try to apply it to everything.

In my personal experience, AI has worked best as a thought partner for product discovery work. Discovery is naturally messy, exploratory, and imperfect, which makes it a good candidate for AI assistance.

A Thought Partner, Not a Replacement

I like to think of working with AI as being similar to working with a capable intern — highly capable but in need of oversight. As smart as it might appear, AI doesn’t really “know” anything.

After all, AI is just a predictive model applied at massive scale to calculate the most likely next word based on patterns in training data. When AI tells you 2+2=4 it’s simply predicting that “4” is the most probable character to follow “2+2=“.

That distinction matters. Because predicting isn’t the same as knowing, and AI has a habit of confidently printing out plausible-sounding completely incorrect information. Your job, as the human operator of a Large Language Model (LLM) is to be the BS detector.

The golden rule is simple: NEVER copy paste AI output.

In the examples below (with the exception of Figma Make), the chat output should be used as an input into other deliverables (i.e. Miro board, flow diagram, wireframe, PRD, etc.).

Think. Refine. Edit. The editing process helps you catch hallucinations while adding your own domain knowledge and strategic context.

Use Case #1: Rough Requirements & User Flows

Skip the blank page and map out high level requirements or user flows to get started.

Think of AI like working with a bright intern: there will be things to fix, errors to correct, and places where the model didn’t quite understand what you were looking for. But, it might get you the first 40% complete on an initial framework by identifying the critical baseline functionality.

The same approach works for exploring “what if” scenarios. With a few prompts, AI can help you think through potential variations or use cases that might add another 10-20% more polish.

Use Case #2: Sample Data & Content

Generate realistic looking prototype content that doesn’t distract so you can spend your energy solving for higher value use cases.

On a recent project, we needed to prototype an input field with a comprehensive listing of construction equipment. We could have spent hours going down this rabbit hole, researching and refining a perfect list. But the important part of our narrative was how the input field fit into a larger flow, not the contents of the field.

As a team, we reminded each other where the value was and asked ChatGPT to unblock us:

Generate a list of 60+ types of construction equipment such as skid steers, dump trucks, and excavators. Define 6-10 categories to group the equipment by type.

This simple prompt gave us realistic data to work with, while keeping our attention on the design challenges that actually mattered.

Use Case #3: Rough Prototyping

Build something that looks and feels “real enough”, without the time or emotional investment of a fully polished experience.

One of the challenges with traditional prototyping, is that by thinking through interactions with the level of detail needed to provide a realistic experience, you end up attached to the direction.

Vibe coding with tools like Figma Make (a comfortable and familiar environment for designers) provides an opportunity to sidestep the risk of emotional investment by dramatically reducing the input required. A prototype that might normally take a month can be put together in less than a week, often off the side of your desk, or while watching Netflix (the prompt responses are slower when generating code, so it’s nice to have a distraction).

Figma Make doesn’t provide the pixel-level polish that a professional designer might create, but it can generate something real enough to garner a reaction, and validate direction with users.

Use Case #4: Brainstorming Partner

Use AI as a sounding board to get unstuck and sharpen ideas, especially when working remote.

When you’re working solo, it can be easy to lose perspective or stall out. AI can act as an always available thought partner — helping you look around corners, stress test ideas, and anticipate reactions before you’re in the room.

Prepping a big presentation for an exec? Ask how a VP might respond to your draft, what questions they’d raise, or what insights they’d bring. It won’t replace real feedback, but assuming you still put in the work, it can sharpen your thinking and surface blind spots.

And when progress starts to lag in a remote environment, a quick prompt can help reset your brain, spark momentum, and keep you moving forward.

The Human Element Remains Critical

AI can accelerate exploration and help you think through problems, but it’s no substitute for the strategic judgement, user empathy, and business context that drive good product decisions. Using it as a tool to generate options and stress test ideas, but keep the critical thinking in human hands.

When used thoughtfully, AI can be a useful tool for UX and discovery work within a human-centered design process. The key is knowing when to lean on it, and when to trust your own expertise.

Cover photo by Nahrizul Kadri on Unsplash