How AI Changed My Design Process After 20 Years
After 20 years of visual design, AI tools cut my iteration cycles from days to hours I use AI for ideation and rough concepts, never for final pixel work The biggest shift: spending less time producing and more time directing AI handles the repetitive parts (resizing, variants, mockups) so I can focus on decisions that matter Your design taste and direction skills matter more now than your production speed The Shift Nobody Talks About I've been doing visual design for over two decades. Started with print, moved to digital, built brands, shipped products. And in the last 18 months, the way I work has changed more than in the previous ten years combined. Not because AI replaced what I do. It didn't. It changed where I spend my attention. Before AI entered my workflow, a typical brand identity project looked like this: research, mood boards, sketch concepts, refine in Figma, build variations, present, iterate, deliver. Each phase took time. Lots of it. The production phases (building variations, resizing assets, creating mockups) ate 40-60% of the total project hours. Now those production phases take a fraction of the time. And that freed-up time goes somewhere more valuable: thinking, directing, and making better decisions. Where AI Fits in My Actual Workflow I don't use AI to generate final designs. I want to be clear about that. The tools aren't there yet for production-quality output in most cases, and even when they are, the result lacks the intentionality that comes from a human making deliberate choices. What AI does well for me falls into three categories. Ideation and exploration. When I'm starting a new project, I use image generation to explore directions quickly. Not to create the final look, but to see 30 possibilities in an hour instead of 5. It's like having a sketching assistant who works in every style simultaneously. Most of what comes back is unusable. But the 10% that sparks something real saves me days of wandering. Production grunt work. Resizing assets across formats. Generating color palette variations. Creating social media templates from a master design. Removing backgrounds from product photos. These tasks used to be mind-numbing hours in Photoshop. Now they're minutes. Tools like Freepik handle a lot of the asset generation and editing that used to eat entire afternoons. Content-adjacent design. Blog headers, social cards, email banners. These need to look good but don't need the same level of craft as a hero section or a product page. AI-generated bases with manual refinement gets me 90% of the way in 20% of the time. The pattern is consistent: AI handles volume and variation. I handle intent and quality. There's a fourth category I experimented with and dropped: AI for layout generation. I tried using AI to suggest page layouts and component arrangements. The results were generic every time. Layouts that looked like every other SaaS landing page. This is where human design sense is irreplaceable. Layout is about hierarchy, rhythm, and guiding attention. AI doesn't understand what deserves emphasis on your specific page for your specific audience. I went back to doing layouts by hand and haven't looked back. What Changed About the Design Process Itself The workflow didn't just get faster. It restructured. Discovery is wider. I explore more directions earlier because the cost of exploring is lower. A concept that would have taken half a day to mock up takes 20 minutes. That means I show clients more options and arrive at better solutions. Feedback cycles are shorter. When a client says "can we try it in blue" or "what if the layout was more minimal," I can show them a rough version in the same meeting instead of scheduling a follow-up. This compresses timelines and reduces miscommunication. The craft moved upstream. I spend more time on brand strategy, user research, and creative direction than I used to. The ratio flipped. Where I once spent 60% producing and 40% thinking, it's now closer to 30% producing and 70% thinking. That's a better ratio for everyone involved. Quality went up, not down. This surprised me. I expected AI-assisted work to feel diluted. Instead, because I have more time to think and less pressure to produce, the work is more considered. Fewer rushed decisions. More intentional choices. On a recent brand project, I spent an extra day on the strategy phase that I would have previously allocated to asset production. The final deliverable was stronger because the foundation was stronger. The client noticed. Revisions dropped. When I explore more directions early and show rough concepts faster, clients feel more involved in the process. They course-correct earlier. By the time I'm in final production, the direction is solid. My average revision count dropped from 3-4 rounds to 1-2. One thing that hasn't changed: I still do all final design work manually. Every pixel in a shipped product goes through my hands. AI is the rough draft, not the final pass. The Skills That Matter More Now If you're a designer watching AI tools evolve and wondering what your role looks like in three years, here's what I've observed. Direction over production. The ability to look at 20 AI-generated concepts and pick the three worth developing is now more valuable than the ability to create those 20 concepts from scratch. Taste, judgment, and the ability to articulate why something works matters more than ever. Systems thinking. Building design systems, token-based approaches, constraint-driven design. These skills compound with AI because they create frameworks that AI can work within. A well-defined design system makes AI output more consistent and useful. Communication. Explaining design decisions to clients and stakeholders is still a human skill. AI can generate options but it can't sit in a meeting and explain why the 16px spacing grid creates better visual rhythm than a 12px one. That conversation is where trust gets built and projects succeed. Technical literacy. Understanding how code works, how designs translate to implementation, and how to bridge the gap between a mockup and a shipped product. AI tools are more effective when you understand their constraints and can prompt them with technical precision. Production speed is becoming table stakes. Every designer has access to the same AI tools. What separates good work from forgettable work is still the same thing it's always been: the quality of the thinking behind it. What I'd Tell Designers Who Haven't Started If you've been avoiding AI tools because they feel like a threat, I get it. I resisted for a while too. Here's what changed my mind: using AI for the boring parts gave me back time for the parts I actually love. Start small. Use AI for one specific task in your next project. Background removal. Color palette generation. Social media asset resizing. See how it feels. Then expand from there. Don't try to replace your entire process overnight. And definitely don't let AI make your creative decisions. Use it like any other tool in your kit: for what it's good at, not for everything. One practical tip: keep a running list of tasks that feel repetitive or low-creativity in your current projects. Those are your AI candidates. Asset resizing, batch exports, mockup generation, background removal, variant creation. Anything that follows a predictable pattern is fair game. Anything that requires judgment, brand context, or user empathy stays with you. The investment pays off faster than you'd expect. I spent about a week integrating AI tools into my workflow properly. Within a month, I had reclaimed roughly 10 hours per week that previously went to production work. That's 10 hours I now spend on creative direction, client strategy, and building products. The designers who will thrive are the ones who pair strong creative judgment with efficient tool use. Not the ones who are fastest at pushing pixels, and not the ones who hand everything to AI and call it done. The middle ground is where the best work happens. After 20 years of design, I thought I knew what my workflow would look like for the rest of my career. I was wrong. And honestly, the new version is better. The Bottom Line AI didn't replace my design skills. It amplified the parts that matter most and automated the parts that didn't. The biggest change is where I spend my time: more on creative direction, less on repetitive production. If you're a designer, the move is to lean into your judgment and taste while letting AI handle volume. The craft isn't dying. It's evolving. And the designers who adapt their process will produce better work, not just faster work.