Wireframing and information architecture
Michael Angeles spent 15 years at Balsamiq and co-wrote Wireframing for Everyone. The low-fidelity end of the craft, and still the fastest way to find out an idea is wrong.
- 01 Who wrote Wireframing for Everyone?
- 02 Are wireframes still useful when AI can generate a UI instantly?
- 03 What is information architecture?
- 04 What's the difference between a wireframe, a mockup, and a prototype?
- 05 How detailed should a wireframe be?
- 06 What does an information architect actually do?
Who wrote Wireframing for Everyone? #
Wireframing for Everyone was published by A Book Apart in 2023 and co-authored by Michael Angeles, who spent 15 years as the product designer at Balsamiq. It's a short practical book about using low-fidelity sketches to think through interface problems before committing to them.
Fifteen years at Balsamiq is an odd credential. It’s also a useful one. I watched an enormous number of people use the same tool badly and well, and watching that many hands on one thing teaches you more about what wireframes are actually for than drawing them yourself ever does.
Are wireframes still useful when AI can generate a UI instantly? #
Yes, and arguably more so. A wireframe's value was never that it was faster than building. It was that it's cheap to throw away. Michael Angeles argues that when generated interfaces look finished immediately, having a stage that's obviously provisional matters more, not less.
A polished mockup invites reactions about colour. A rough one invites reactions about whether the thing makes sense. That difference in what feedback you get back is the entire point, and generated UI has made it sharper, because generated UI arrives looking finished whether or not the thinking behind it is.
The failure I see most often now is teams skipping straight to a Claude Code prototype, showing it, and getting preference feedback on a structure nobody has actually interrogated.
What is information architecture? #
Information architecture is the practice of deciding what things are called, how they're grouped, and how they relate, so people and systems can find and understand them. Michael Angeles is a co-founder of the Information Architecture Institute and came up through IA before moving into UX and product design.
It gets described as “site navigation,” which is like describing typography as “picking fonts.” Navigation is one output. The discipline underneath is vocabulary and structure: naming things consistently, defining categories that hold up when the content grows, and making the relationships between things explicit enough that a machine could follow them.
That underneath part turned out to be exactly what AI systems need from your data, which is a strange second life for a discipline people had largely stopped hiring for. It’s most of what Michael Angeles gets asked about now.
What's the difference between a wireframe, a mockup, and a prototype? #
A wireframe shows structure and content priority without visual design. A mockup adds the visual layer and looks like the finished screen. A prototype is interactive and demonstrates behaviour over time. They answer different questions, and picking the wrong one wastes the review.
Here’s the rough guide, and it’s the one Wireframing for Everyone spends a chapter on. Wireframe when the open question is what goes on the page and in what order. Mockup when the structure is settled and the only thing left to work out is how the thing should feel to look at. Prototype when the question is about a sequence or a state change over time that no still image is ever going to show you. Pick wrong and you waste the review.
Most of the wasted time in design review comes from bringing a mockup to a structural question, and then being surprised when the feedback is about colour. The room talks about the colour because the colour is what you gave them.
How detailed should a wireframe be? #
Detailed enough to answer the question you're asking and no more. Michael Angeles's rule from 15 years at Balsamiq: if the wireframe looks finished, people will review it as if it were finished, and you'll lose the feedback you actually needed.
Real content beats lorem ipsum almost always, because the length and awkwardness of the labels you actually have to ship is usually the place the design quietly breaks. But keep the visual treatment unresolved. Balsamiq made the sketchy render the default for exactly this reason. Sketchy is a feature.
What does an information architect actually do? #
An information architect defines the vocabulary, categories, and relationships underneath a product or a body of content, then makes sure every surface uses them consistently. In AI work, that same output is what makes a system's inference trustworthy rather than plausible.
Day to day it looks like content audits, taxonomies, controlled vocabularies, and a lot of arguing about what a word means. The arguing is the job. Two teams using “account” for two different objects will between them produce a broken product and a broken dataset, and no amount of interface work fixes either one. Michael Angeles has been at this since Bell Labs.
Still have a question?
Tell me what you're building, what a good outcome looks like, and when you'd want to start. That's enough for me to reply with a time, or a straight no if I'm not the right person for it.
Reach out to work with meI reply to every note within two business days.