The world is changing fast, and right now, almost every designer is racing not to get left behind in the AI wave. Many of us feel overwhelmed by the current shift driven by AI products. Some designers believe there’s already more than enough to learn and upskill on without adding AI to the mix.
Since the beginning of this year, I’ve explored building for AI: the nuances, best practices, and misconceptions. I’ve validated some insights while also demystifying incorrect assumptions through the design of an AI-powered product I’ve been working on.
To make this easier for designers—whether you're just learning about AI, about to start building an AI product, or in the thick of the design process—this edition of The Designers Edge newsletter is a no-nonsense guide to designing and building for AI.
Founders who want to build AI products should read this too. I won’t speak in design jargon, just plain, everyday language so you can understand and experience what it means to build for AI through this article.
TL;DR Checklist for Designing AI Products (If you don’t have the patience for a long read)
Designing for AI goes beyond the screen: it’s about human-AI interaction.
Think of AI as a collaborator, not just a backend tool.
Collaborate early with AI engineers: align on capabilities and limitations.
Research is essential: solve real problems, not just follow trends.
Feedback from real users is the most valuable form of validation.
Define your AI’s tone and personality clearly: Users trust transparency.
Build for trust: don’t pretend your AI is perfect.
Set clear boundaries and expectations for AI behaviour.
Provide immediate, clear feedback throughout the user journey.
Design for user autonomy and emotional resonance.
Think about non-visual elements: motion, sound, and haptics.
AI is a design opportunity, not a threat to your job.
Misconceptions: AI doesn’t need to be human; it needs to be trustworthy.
AI is Reshaping the World
From how we communicate to how we learn and work, AI is changing everything. One thing is clear though: building AI products is not just about designing interfaces. It’s about shaping how humans interact with intelligent systems. You’re not just designing screens; you’re helping to change the world.
This goes beyond visuals, usability, code structure, or scalability. It’s about behavior, trust, intent, and most importantly, collaboration.
Here are a few things to consider when designing AI products:
You're Not Just Designing for Users, You're Designing for Relationships
When designing AI products, you must expand your focus beyond the user or the business. You also need to consider the dynamic between the user and the AI system. Think of it as designing a companion. You’d consider the user’s preferences, the type of companion they need, and how that companion should behave. With traditional products, the focus is on how the user interacts with the interface. With AI, you must also design how the AI interacts with the user.
This means mapping out the emotional arc of the user’s interaction with the AI. What does trust look like? What does comfort feel like? What are the AI’s capabilities? These are the questions that lead into the next point.
Collaborate Early and Often with Engineers
Designing for AI means working closely with engineers: more closely than you may be used to. Your understanding of AI will be limited, and you can’t predict how the system will behave without their input.
You need to understand:
What the AI model can do
Where its limitations lie
How it can be fine-tuned or personalized
In my case, I collaborated with the AI engineer to define what our AI could and couldn’t do. This shaped everything—from brand personality to tone of voice, user flows, and even error handling.
Treat your AI engineer like a co-designer.
Share flows early. Align on capabilities. Define behaviours together.
Research is Your Compass
The more novel your AI solution is, the more research you need to do. Without research, you're flying blind.
Research existing AI solutions
Review traditional alternatives
Look at user data
Read across platforms
Dive into forums, comments, reviews
Approach your product with a problem-first mindset. Many teams today plug AI into their products simply to ride the hype, without real justification. Often, these additions are unnecessary. If you had asked, “What problem are we trying to solve?” before building, you may have realised that AI wasn’t even needed. Not every problem requires AI. Sometimes users just want to wash their hands with soap and water, not use a washing machine.
Feedback is Your Single Source of Truth
We’re still early in the AI era. Discoveries are happening fast, and best practices are evolving. In this environment, assumptions are dangerous. The best way to know how your product performs is to put it in the hands of real users quickly.
Research and collaboration will guide you, but feedback is your telescope.
I’ve learnt not to overemphasise internal testing. Let developers ship fast. Once QA is done, get it into users’ hands—even in beta. Because nothing compares to feedback from a live app.
Every AI Has a Personality. Define it.
Not every AI needs a persona. In fact, anthropomorphic (human-like) AI may be on its way out. Users are increasingly wary of systems that try too hard to be human. Instead, we’re seeing a rise in hybrid AIs: tools with clear, non-human identities but strong, defined personalities. Think ChatGPT.
Give your AI a consistent tone and personality that builds trust. Avoid fake emotional mimicry. Be transparent. Users can handle it.
Define how your AI responds to uncertainty, admits limitations, and builds rapport.
Trust is Paramount
Transparency is the foundation of trust. Don’t present your AI as all-knowing: it’s not. Expectations vary wildly between users. Some are sceptical. Others are optimistic. Either way, users are still adjusting to AI. Mistakes are okay, but how you handle them is crucial.
Use error-prevention techniques and consider edge cases. Build an experience that feels safe, even when the AI is imperfect.
Set Boundaries and Expectations
Managing user expectations is one of the hardest parts of AI design.
Is the AI always right?
How reliable is it?
Be explicit about what the AI can and can’t do. Use microcopy, interface cues, and feedback animations to reinforce this. Reduce surprises. Increase confidence.
Feedback is Twice as Important in AI Design
Unlike traditional products, AI products demand real-time feedback. Users expect immediacy: they associate AI with speed. Your design must meet those expectations.
Design your product to provide clear, immediate feedback, especially when AI is processing or responding.
Design for Emotion, Not Just Tasks
Identify emotional friction points in your flow. Use visuals, tone, and interactions to soothe, reassure, and empower. Your product should feel like a safe space.
Design to Give Users Autonomy
The best way to build trust is to give users control. Let them personalise their experience. Don’t box them in: design options that give them freedom and flexibility. This reduces frustration and keeps expectations in check.
Think Beyond the Screen
AI is not just an app: it’s an experience. Think about how your AI feels, sounds, and responds. Consider sound, motion, microinteractions, and even haptics to bring it to life.
Most Designers Worry AI Will Replace Them. Few See the Opportunity.
Instead of waiting for AI to take your job, why not design the AI of today?
Common Misconceptions Designers Have About AI
"It has to feel human." Wrong. It has to feel trustworthy, not human.
"I only design the UI." You design the behaviour, the feedback, the guardrails, and the emotional flow.
"AI is just a backend feature." AI is a co-actor in your user’s journey. It is the product experience.
Final Thoughts
Designing for AI is a thrilling challenge. It pulls you into strategy, language, systems thinking, human psychology, and deep engineering collaboration. It invites you to think bigger: not just about interfaces, but about interactions.
If you're a designer stepping into this space, remember: You’re not just designing screens. You’re designing relationships between people and machines.
I’m Evergreen Odeh, a senior product designer, business strategist, and mentor. With 4+ years of experience across e-commerce, gaming, AI, edtech, and more, I bring deep product thinking and emotional nuance to every product I design. If you’re building something AI-powered, let’s talk.
You’re reading this because you subscribed to The Designers Edge. If you’re not yet a subscriber, join here.
I’m building this into a thoughtful space for designers and founders who want to create with confidence, clarity, and purpose. I’d love to have you on the journey.
