Skip to content
AppDues
Log in
Contact salesSign up

App Icon Design: Breathing Room, AI References, and the Glass Frame

An app icon has one job before anyone opens the app: make the product feel worth keeping in the dock. That is a design decision, not a decoration. This guide turns a working designer's process into a checklist you can use when you commission, sketch, or review an icon — including what to do when a client arrives with a folder of AI-generated options, and how to scope a commission that is short by nature.

Start with the shape, not the color

The useful sequence is deliberately unglamorous. Sketch a batch of concepts on paper, pick the strongest silhouette, build it in a vector tool, refine it in a UI design tool, and only then add texture and color. Texture belongs last because it hides weak shapes; color belongs last because it cannot rescue an unclear one.

A practical brief should name the symbol, the intended scale, and the platform frame before anyone opens a design file. If those three are unsettled, the review will drift into taste arguments.

The most common icon mistake is scale

The single most frequent fix is to make the symbol smaller. Designers tend to enlarge a logo until it fills the squircle, which usually reads as cramped rather than confident. Give the mark breathing room; even a generous grid circle can be slightly too large for a comfortable result.

A quick review test: view the icon at dock size and at a small list size. If the symbol touches the edge or the negative space disappears, shrink it before changing anything else.

Recognition over literalism, and where the craft comes from

The strongest icons are not always literal pictures of the app. A recognizable object can do more work than a diagram of the feature set, because the icon's job is to make the product feel desirable in the dock. The Tide Guide icon, designed by Matthew Skiles, was shown on camera in an Apple WWDC video — a useful reminder that recognition and craft travel further than a literal label.

That craft is often built outside formal training. Skiles grew up on New Zealand's South Island and taught himself design; at 15 he worked through an HTML book and, with his brother, built a one-page site for a tennis coach. The pair then shipped their own iPhone game, Rock'n Rockets, and Skiles drew its first app icon — a title MacStories covered twice and roughly a hundred people played. Paid icon commissions began around 2014, and they did not arrive through a portfolio, since he had none. They came from people who already knew his work on other projects and trusted him enough to try him out. The follow-through was disciplined: he published icon work only, never web work, on Dribbble first and Twitter later, until Christian Selig hired him for Apollo's alternate icons and exposed his name to a far bigger audience.

The lesson for a first commission is not that credentials are useless, but that a focused body of work and a few people who trust you can substitute for a formal portfolio. When reviewing a concept, ask: would someone want this object on their home screen every day? If the answer is no, the symbol is probably too descriptive and not distinctive enough.

Using AI references without losing the decision

AI images can be a genuinely useful communication tool. A generated reference can show a client's intended mood faster than a long written brief. The failure mode is different: infinite generation creates choice paralysis. A client may keep producing more options after a direction has been recommended, and the project stalls.

The worse version is when a client falls in love with one generated image and asks for it to be reproduced. At that point the designer is redrawing someone else's output rather than solving the icon problem. A clear brief should state whether AI is being used as a reference, as a mood board, or as final artwork — and who makes the final call.

There is also a reasonable middle position: deliberate patience. Tools tend to get easier to adopt over time, so watching first carries little risk; the same wait-and-see approach once applied to a now-standard design tool ended in adoption once the tool proved itself. The one area that pulls a working designer in sooner is custom generators for textures and assets, where the designer keeps control over what is being generated rather than handing the decision to the model.

Platform frames: advised versus forced

Apple's icon conventions have moved through distinct eras. The flattening after iOS 7 removed texture as a hiding place and pushed designers toward stronger shapes. The Big Sur era treated the squircle as a convention that could be broken, which tends to produce better results than a hard rule.

The Tahoe-era glass frame is a harder constraint. Icon Composer can save time when a client wants a system-native look across platforms, but the mandatory edge highlight is not optional. A designer can still put glass inside the icon; the frame itself is the part that cannot be removed. Plan for that constraint early rather than treating it as a final polish step.

Scoping the commission: short work, steady demand

Icon work is an unusual consulting business. Engagements are short, repeat work is limited, and there is a constant need for the next client. A designer can acknowledge that plainly and still keep going, because the work is chosen for the enjoyment of the craft rather than for a clever business model — and because it keeps working.

What keeps it working is the scale of the App Store itself. Even after a career's worth of icons, there is always a new developer arriving with a new app and a request for an icon. That is a job that did not exist a generation ago; it exists because the platform made icon appearance something developers are expected to care about.

For a commissioner, that shape has practical consequences. Scope one icon or a small set rather than an open-ended exploration, agree the number of review rounds up front, and treat the brief as the thing that protects both sides from drift. For a designer, it argues for keeping a visible, focused body of icon work so the next client can find you without a formal portfolio.

A short decision checklist

Use this before approving an icon:

  • Is the symbol smaller than the frame, with visible breathing room?
  • Does the silhouette read at dock size and at list size?
  • Is the symbol a recognizable object rather than a literal feature diagram?
  • Was color chosen after the shape was settled?
  • If AI was used, was it a reference, a mood board, or final artwork — and who decided?
  • Does the icon account for the platform's mandatory frame, not just the inner artwork?
  • Is the number of review rounds and the scope of the commission written down?
  • Would the intended user want this object on their home screen every day?

Frequently asked questions

What is the most common app icon design mistake?

Making the symbol too large. A mark that fills the squircle usually looks cramped; shrinking it slightly and leaving breathing room tends to improve the result.

Should an app icon be a literal picture of what the app does?

Not necessarily. A recognizable object can be more effective than a literal feature diagram, because the icon's job is to make the product feel worth keeping in the dock.

Can AI be used in an app icon workflow?

Yes, as a communication tool. A generated reference can convey a client's intended direction faster than a written brief. The risk is choice paralysis from endless generation, or a client asking for a specific AI image to be reproduced rather than solved. A deliberate wait-and-see stance is also reasonable, with custom generators for textures and assets being the area where designer control matters most.

What is the glass frame in Apple's Tahoe-era icons?

It is a mandatory edge treatment applied to the icon frame. Designers can place glass inside the icon, but the frame highlight itself is not optional. Icon Composer can help produce a system-native look across platforms.

Do you need a portfolio to get app icon work?

Not necessarily a formal one. Work can arrive through people who already know your output from other projects. Keeping a focused, visible body of icon work — rather than a mixed portfolio — is what lets a wider audience find you later.