As designers, we often hear that AI is here to replace us. I take a different view: AI is here to remove the friction between the idea in my head and the product in the user’s hands.
I’ve been auditing my own workflows to see where AI can act as a force multiplier. I’m not looking for a machine to do the thinking for me; I’m looking for leverage.
Here are five practical ways I’ve successfully integrated AI into my design process to work faster, think broader, and ship higher-quality work.
1. Information Architecture
One of the most time-consuming parts of early-stage discovery is making sense of unstructured data. I’ve found AI to be an incredible partner for clustering and labeling.
When dealing with complex navigation systems or feature sets, I use AI as a viable substitute for preliminary open card sorting. By feeding an LLM a list of features or content, I can ask it to propose logical groupings based on industry standards or specific user mental models.
The Caveat: AI is a great starting line, but not the finish line. While it generates excellent hypotheses for grouping, these architectures still require rigorous validation with actual end-users to ensure they hold up in reality.
2. UX Writing
We’ve all stared at a blank canvas trying to write the perfect error message that is helpful, concise, and human. AI has become my go-to for unblocking this process.
I use it to generate variations for:
- Error messages: Turning “System Error 404” into helpful guidance.
- Transactional Emails: Drafting copy that informs without overwhelming.
- Microcopy: Brainstorming button labels and section headings.
Pro Tip: The output is only as good as the input. I always prompt the AI with examples from our existing documentation or style guide to ensure the tone matches our brand voice. This consistency is key to maintaining a seamless user experience.
3. Brainstorming
It is normal to feel stuck in a design box. When you focus on the same problem space for months (or years) using the same component library, your creative muscles can atrophy. You start designing on autopilot.
I use tools like Figma Make or Lovable to break out of these ruts. I might prompt them to “generate a dashboard of KPIs for admin users” or “streamline this document management system with more salient calls-to-action.”
You might not hand these results directly to engineering, but that isn’t the point. The goal is to:
- Explore avenues you hadn’t considered.
- Remind yourself of concepts you’ve long since forgotten.
- Generate divergent ideas so you can focus on practical solutions.
Note: Always filter these suggestions through the lens of your roadmap. Just because AI suggests a flashy new interaction doesn’t mean it fits the current scope or engineering bandwidth.
4. Prototyping (The Shift to Code)
I have almost entirely abandoned traditional point-and-click prototyping tools.
Early in my career, they served a purpose. But lately, connecting endless frames and managing Figma’s component states feels like “the juice wasn’t worth the squeeze.” Often, a click-through prototype provides little value outside of the specific meeting where it is demoed.
The AI Advantage: Instead, I use AI to help me write actual code for my demos. The value here is immense.
- Realism: Executives and stakeholders are far more impressed when I demo code from a local branch rather than a mock-up.
- Comparison: I can switch between branches in real-time to showcase competing design alternatives.
- Longevity: The work done here isn’t thrown away; it’s a foundational step toward the final build.
5. Component Development
This is perhaps the biggest value add for a modern Product Designer. The question is: What is more valuable to an organization? A person who can generate mockups, or a person who can generate mockups and help make them a reality?
I use AI coding assistants to build out the components I design. However, I never just write code and “chuck it over the fence” to engineering asking for a PR review. I thoughtfully review the output, creating a bridge between design and development.
By doing this, I increase my own technical domain knowledge. This leads to:
- Better alignment between design vision and the shipped product.
- Faster shipping of features.
- More bugs squashed before they reach QA.
The Bottom Line
Integrating AI isn’t about cutting corners; it’s about allocating your brainpower to the problems that actually require human empathy and strategic thinking. By automating the structural and repetitive parts of the process, I can focus on what matters most: solving user problems.