"I never use Figma anymore. Now I use [insert new AI tool]."
It’s the conversation happening in product design circles right now. And it’s definitely getting people’s attention.
Often, the reaction that quickly follows is, "If I don’t also abandon that tool and adopt this new one, I’m going to be left behind."
I don't have an issue with people sharing what's working for them. I have an issue with the implication that follows — that the rest of us should be doing the same thing. The skill product teams need right now, maybe more than ever, is the judgement to know which tool the task in front of them actually calls for. That takes both the curiosity to try new things in addition to the experience to know what you're solving for with the tool.
Different problems, different workflows
We work with companies coming at AI from every angle, from teams with an established stack that already works for them to teams who know they need AI but don't know where to start. Regardless of stage, the teams we see having the most success, and the products with the happiest users, are the ones who've accepted that no single tool solves every problem they have.
That's true inside Innovatemap too. In one engagement, I'm designing almost exclusively in Figma. In another, I'm moving between multiple AI design and coding tools and never open it. Neither workflow is better or more effective.
It might be faster to execute a task with AI, but speed isn't always the goal.
Recently, a client handed us a brand-new feature with almost no requirements defined. Rather than open a blank Figma file, I used Claude to think through the problem itself, sketching visual directions while working through the interactions, edge cases, and error states a user might actually hit. By the time I landed on a direction, I was confident in the design. More importantly, confident that the requirements underneath it actually held up. That's when I moved into Figma to build it for real.
What's in my rotation right now
The specific tools change. The logic behind picking them doesn't. Here's what that looks like for me right now:
- Figma — Handles high-fidelity design, especially work tied to a design system that already exists there.
- Claude Design — Where complex prototyping happens: understanding interactions and flows, or moving fast through early exploration.
- Codex – For quicker design and prototyping brainstorming and exploration.
- Claude Code — Takes on the heavy lifting of a design system migration, confirming components and tokens map correctly from the old system to the new one.
- Cursor — Built for smaller, contained jobs. Like recently, a quick Figma plugin to solve a one-off problem.
What doesn't change when the tools do
What stays the same across all of the workflows is the judgment behind them. It's knowing when a client conversation needs to happen before a pixel gets pushed, when something becomes a documented pattern instead of a one-off, when a design feels wrong before I can say why.
None of those were ever a speed problem.
AI can shorten how long it takes me to get to an answer. It can't make the call for me.
Ask what the task needs, not what's new
Forget which tool is newest, or which one everyone just switched to. Ask what the task in front of you actually needs. That question only has a good answer if you already know what the task is and what a good outcome looks like, tool aside. That's the same judgment I used before any of these tools existed. It's the same judgment that decides what I reach for now.
Kristin Bailey makes a great point in "The Shovel Doesn't Know Where to Dig", AI is a capability, not a strategy, so it only works once you know the problem you're pointing it at.
Katie Lukes took the question further in "Where Design Expertise Ends and AI Tools Begin." Once you're inside a project, where does AI's output stop and your expertise need to take over? This is the question that comes one step earlier. Which tool do you pick up in the first place, and why?
Neither question has a permanent answer. The tools will keep changing. The judgment won't.
If you're figuring out where AI actually fits into how your product team works, and where it doesn't, our product experience team would love to help you work through it.
