Hola Smashing Friends,
In times when we can build anything with AI, we need to clearly understand what we actually want to build. It’s not the idea, but rather the context around it — from problems to solve and features to design, all the way to how we build confidence and trust.
AI can’t resolve years of accumulated debt in code, design or content. If anything, it only surfaces underlying inconsistencies and amplifies them in its responses. To prevent it from happening, we need to apply a healthy dose of critical thinking. And that’s what this newsletter is all about.
Let’s dive into how to design for human oversight, how to establish design principles, how to design for trust — and for different modes of AI prompting.
Just last week, we wrapped up our lovely first SmashingConf in Amsterdam, and it was a truly magical and lovely experience with 450 people in attendance.
If you missed it, join SmashingConf Antwerp 2026 🍫 — focused all around UX, design, research, complex UIs and design patterns for AI. With waffles, chocolates and quite a bit of UX. Jump to all topics →
Also, here are a few upcoming online workshops, and some of them are launching this week:
Thanks for being smashing, everyone!
— Vitaly
1. Four Modes Of AI Prompting
What are you actually trying to get done? That’s the crucial question we need to consider when we want to work smarter with AI tools. To help us address it, Tey Bannerman created the Four Modes framework. Whether it’s analyzing competitor activity, brainstorming ideas, or checking if a strategy actually holds together, the framework provides principles, prompts, and examples to effectively tackle different scenarios with the help of AI.
2. Designing For Trust
Trust in AI is a tricky topic. Some people are so skeptical of AI that they refuse to use it even when it could be genuinely helpful, while others follow AI recommendations on auto-pilot. So, how can we design for trust, without pushing people in either direction? According to Kristine K., the keyword is “calibrated trust.”
If you want to dive deeper into how to design, build, calibrate, and maintain user trust, be sure to also check out Vitaly’s free lesson on Maven and the accompanying Google doc with lots of resources for designing for trust and confidence in AI.
3. The Atlas Of AI Interaction
How can we create AI systems that are genuinely human-centered? The AI Interaction Atlas explores over 100 patterns, visual elements, and reusable components for crafting effective human-AI interactions. By giving teams a shared vocabulary to map roles, responsibilities, and decision points within AI systems, it helps uncover gaps and clarify responsibilities when designing complex, multi-step AI experiences.
From our sponsor
DatoCMS: The Headless CMS Picked By Devs, And Weirdly Enough, Loved By Editors
You know that feeling when a tool just... works? GraphQL and REST out of the box, smooth content modelling, CLI, Visual Editing, cache tags, plugins, and so much more. And the editor side is just as solid. No wonder Jeff called us “unbelievable”. DX-first, editor-approved. Try DatoCMS for free →
4. UX And Product Design Challenges
Working through a design challenge on your own before turning to AI to quickly generate ideas and concepts sharpens your critical point of view and your design skills. If you want to stretch your problem-solving muscles a bit — or prepare for your next UX interview — Vitaly compiled UX challenges and product design exercises, as well as tips for working through them.
5. Practical Guide To Design Principles
“Design should not be a matter of taste; it must be guided by our goals and values.” Design principles help us achieve exactly that. As Vitaly argues, they avoid endless discussions and rally a team around a shared purpose. In his practical guide to design principles, he dives deeper into why design principles matter and how to run a workshop to establish your own set of principles.
6. Five Levels Of Context Engineering
Context is key to reducing the hallucinations and broad assumptions that AI is known for. So, how can we “onboard” AI in a way that gives it just enough context to tackle a task effectively? Matthew Alverson suggests five levels of context engineering to help construct a model’s context window and thereby enhance its output.
From our friends
beyond tellerrand Düsseldorf 2026: Fifteen Years Of Bringing Curious People Together
One track. 500+ attendees. Practical talks from designers and developers who care about doing great work. Come for the speakers, stay for the conversations. Tickets are limited — get your ticket!
7. Designing Human-AI Oversight
Whether it’s job applications that get rejected by algorithms before recruiters ever see promising candidates or customer service bots that frustrate instead of giving agents the context to actually solve problems, a lot of AI implementations fail at giving humans the authority, time, and understanding they need to be genuinely effective. To help us do better, Tey Bannerman shares a framework for designing human-AI oversight that actually works.
8. Accessible UX Research, Now Shipping 📚
We’ve got exciting news! Our newest Smashing book, Accessible UX Research by Michele A. Williams, is finally shipping worldwide! Get the book right away or order the eBook for instant download.
Meet our brand-new book: “Accessible UX Research” by Michele A. Williams. Printed copies shipping now.
Accessible UX Research is your practical guide to making UX research more inclusive of participants with different needs — from planning and recruiting to facilitation, asking better questions, avoiding bias, and building trust. Download a free sample (PDF, 2.3MB) or get the book right away.
9. Upcoming Workshops and Conferences
That’s right! We run online workshops on frontend and design, be it accessibility, performance, or design patterns. In fact, we have a couple of workshops coming up soon, and we thought that, you know, you might want to join in as well.
With online workshops, we aim to give you the same experience and access to experts as in an in-person workshop from wherever you are.
As always, here’s a quick overview:
10. Person Of The Week: Dennis Lenard
Please give a warm round of applause for our Person of the Week: Dennis Lenard. Dennis is a partner at Creative Navy, a design studio specializing in medical software, embedded systems, and high-stakes technical applications. He lectures about ergonomic UX/UI design at King’s College and has published the first textbook on medical software.
Dennis has 20 years of experience and focuses on improving usability, structuring complex systems, and aligning product design with business goals. He also shares his knowledge as a speaker.
Thank you for everything you do for the community, dear Dennis! 🧡
11. Recent Smashing Articles
That’s All, Folks!
Thank you so much for reading and for your support in helping us keep the web dev and design community strong with our newsletter. See you next time!
This newsletter issue was written and edited by Cosima Mielke, Vitaly Friedman, and Iris Lješnjanin.
Sent to truly smashing readers via Mailchimp.
We sincerely appreciate your kind support. You rock.
Follow us on Mastadon • Join us on Bluesky
