Designing in the Age of AI

In September 2026, the Interaction Design Foundation (IxDF) ran a workshop along with BugHerd: The AI-Assisted Designer: Faster Delivery Without Losing Creative Control. 

Hosted by IxDF’s Edward Rodriguez, the workshop was presented by BugHerd’s Head of Product, Richard O’Brien, and CEO, Stephen Neville.  Together, they covered the designer's role in the age of AI, how the design process is changing, and how to take the pain out of client reviews. It also included a live demo of AI fixing website feedback, and finished with a Q&A. Here are the main takeaways.

‍

Don't believe the hype

Scroll through LinkedIn for five minutes and someone will tell you that you're already behind. Their workflow is the right one, their tools are the ones to use, and if you're not working exactly like them, you're failing. It's easy to feel like everyone else has it figured out.

But design teams everywhere are working through the same questions you are: which tools are worth learning, how much AI should do, and how to fit it into real client work. There’s no single perfect way to work with AI yet. The tools change every few months, and what works for one team can be a poor fit for another.

Most people are learning as they go, so don't chase every trend or copy someone's setup just because it looks impressive online. Instead:

• Keep trying things. Test new tools on small, low-risk tasks before using them on client work.
• Keep learning. Notice what works and what doesn't, and share it with your team.
• Keep adjusting. Your process will keep changing, and small improvements add up over time.

Progress with AI comes from steady experimenting, not from getting it perfect on the first try.

‍

Your experience matters more now, not less

AI is a predictive machine, so by default it drifts towards ‘the average’. Ask it for a design and you'll often get the same look: 

• lots of gradients 
• too much bold text 
• little eyebrow headings everywhere
• and plenty of purple

A designer's job is the opposite - it's to help clients stand out. The quality of what AI produces depends heavily on the expertise of the person guiding it. Your taste, judgement and years of experience are what turn average AI output into great work. 

People who know their field get far more out of AI than people who don't. For newer designers, that makes building real design knowledge, and finding mentors, more important than ever.

‍

Design and development are merging

The line between designers and developers is fading fast. According to Figma research, 60% of developers now take part in design work, and 40% of designers are getting involved in coding. That second number has doubled in the past year.

The old web design process went in a straight line: 

That is now collapsing into one loop. 

The pace is striking! BugHerd's web development team gets roughly twice as much done as they did 12 months ago with the same people, and the engineers now write very little code by hand. They mostly direct AI agents instead.

‍

Share in code, not mockups

This was the biggest idea of the session: stop sharing static mockups and start sharing a real, working version of the site as early as you can.

The closer your work is to the finished product, the less guesswork clients have to do. We've all had a client approve a design, then see the built site and say, "Oh, I didn't realize it would look like that on mobile." A clickable, working page with real text brings that feedback out early, when it's easy to fix.

Code is cheap and quick to make now. AI agents make much more sense of code than of a mockup, and developers can pick up right where you left off instead of rebuilding your design from scratch. Getting a working site online is easy too: with tools like Cloudflare or Webflow, it's often just a click or two.

‍

Think in systems, not one-off deliverables

Instead of designing each piece from scratch, ask yourself: "How can I make sure I never have to do this again, or that an AI agent can do most of it for me next time?"

This frees you up for the creative work that actually makes a difference and it also makes you more employable. The person who automates parts of their own job is exactly who companies want to hire, because they're curious, they try new things and they ask the right questions.

“The person who automates parts of their own job is
exactly who companies want to hire”

‍

Use AI to explore ideas, then refine them yourself

AI is most useful at the start of a project, when you're facing a blank page and want to explore lots of directions quickly. But getting to polished, final work still needs a designer's hands, often back in Figma.

Practical tips:

• Work in small sections. Ask AI to explore one part of a page, like a hero area or a dashboard widget, rather than a full page. You'll get more variety and use fewer tokens.
• Bring it back into Figma. Screenshot an idea you like, have Figma turn it into an editable design, and keep refining from there.
• Use cheaper models for the doing, expensive ones for the reviewing. Treat cheaper AI models like junior designers who do the work, and a top-tier model like a creative director who reviews it. As AI companies push to become profitable, managing token costs will matter more and more.
• Learn to prompt for creativity. Getting AI past its default style is a skill. For a great place to start, read this guide in Lenny's Newsletter, where a former Apple designer shares the prompting techniques he uses to get original, high-quality designs out of AI.

‍

Learning GitHub is the key

GitHub scares off many designers because of its technical words and seemingly complicated terminal, but it's simpler than it appears to be.

You don't need the GitHub terminal either. Desktop apps like Tower let you do all of this through a normal visual interface.

‍

GitHub Learning Resources

• Git for Design Teams
• GitHub for Beginners: Step-by-Step Guide
• Tower Git client for Mac & Windows
• GitHub for Designers: What You Need to Know to Work with AI (you need to be an IxDF member to access this masterclass)

‍

Fix the review cycle

Sharing in code creates a new problem: How do you collect feedback? In Figma, clients just click and comment. On a live website, feedback often turns into a mess of screenshots, spreadsheets and email chains. Teams might budget about 10% of a project for reviews, but this stage is generally heavily underestimated, and usually takes three times longer when using traditional methods.

To speed the review process up, use a feedback tool, such as BugHerd, that makes commenting on a website easy for clients. They should be able to leave feedback by simply pointing, clicking, and leaving a comment, with no login required. 

Each piece of feedback should automatically capture a screenshot, along with the reviewer’s technical details - everything that designers and devs need. The top tools also have an inbuilt task board, and AI features to help web teams resolve issues even faster, as well as integrating with PM tools.

It’s estimated that teams using website feedback tools with AI features are able to automate around 40% of their feedback work.

‍

Highlights from the Q&A

What is an "AI-native" designer? Someone who is comfortable working with AI every day and doesn't avoid it. They understand their job is about delivering better outcomes for clients and customers and will use AI to achieve this.

How do you make sure AI output isn't rubbish? Give clear briefs and show the AI what ‘good’ looks like. When the output doesn’t hit the mark, explain exactly why, then save that lesson as a reusable instruction so that your AI tool improves next time. Screenshots of mood boards or examples help a lot. It also helps to manage how AI uses your design system. A tool like uirules.com lets you set your design system up once and share it with all your AI agents, so every design they produce sticks to the same colours, fonts and components.

How do you avoid burning through tokens? Save the expensive top-tier models for decisions and use cheaper models for the hands-on work. Cheaper models are now good at writing code.

Are companies replacing designers with AI? There is still plenty of design work. And if everyone uses AI with no designer involved, every company’s branding and creative assets will end up looking the same.

‍

Final thoughts

The way designers work is changing fast, and it can be hard to keep up. New tools appear constantly, and it's easy to worry that your skills are becoming less useful. But it’s important to remember that your experience is worth more now, not less.

AI is only as good as the person guiding it. It can't tell good work from average work on its own, but you can. If you know what a great layout looks like, how users behave and what clients actually need, you can spot when AI gets it wrong and steer it towards something better. Someone without that knowledge will simply accept whatever it gives them.

So the years you've spent building your eye for design aren't wasted. They're what makes AI genuinely useful in your hands, so learn to work with these tools, teach them what good looks like, and you'll get far more done than you could before.

One last piece of advice: look after your mental health. When you can suddenly do ten things at once, it's easy to burn out. Most people outside the design industry aren't thinking about AI at this level, so go easy on yourself, help each other out, and keep learning one step at a time.

‍

Watch the Recording

‍

Resources

AI-Assisted Designer: Workshop Takeaway Sheet 

How to turn your AI into a world-class designer (Lenny’s Newsletter)

State of the Designer 2026: Designers are leaning into the messy middle

BugHerd Website Feedback Tool - free trial

‍

About IxDF

The Interaction Design Foundation (IxDF) is a global leader in online design education, with over 1.1 million course enrolments. Members get access to more than 40 self-paced courses covering everything from UX basics to advanced topics for experienced designers. Membership also includes job matching, free design tools and a worldwide community, with forums and local groups in over 100 countries.

‍

About BugHerd

BugHerd is a website feedback tool used by thousands of design and development teams around the world. Clients and teams simply point, click and comment on a live website, with no login needed. Every piece of feedback automatically captures a screenshot and the technical details developers need, and lands on a built-in task board. BugHerd also connects with AI agents, so routine fixes can be handled automatically and your team can focus on the work that matters. BugHerd is based in Melbourne, Australia.