Your resume is read by software twice, not once
Most advice about “clean resume design” still targets a single filter: the applicant tracking system that scans for keywords. That filter still exists. SHRM’s 2026 State of AI in HR report found that over 90% of large employers run some form of ATS before a human ever opens the file. But many of those systems now add a second filter on top: a language model that reads the parsed text for evidence, not just keyword matches. Harvard Business Review covered this shift in June 2026: generative AI is changing which resume signals actually predict a strong hire.
A minimalist resume has to pass both. The design choices below aren’t about looking modern. Each one exists because it changes what the parser extracts or what the AI screener reads.
Why multi-column layouts and graphics break the parse
Single-column layout, standard section headers, and no text boxes or embedded graphics aren’t aesthetic preferences. They’re how you avoid the most common parsing failure: an ATS reading a two-column resume top-to-bottom across both columns instead of down each one, scrambling your work history into the wrong order or dropping lines entirely.
This is also where a lot of Canva templates fail. Canva makes it easy to design something that looks sharp on screen and still reads as garbage to a parser, because the visual layout (columns, icons, text boxes) doesn’t map to the linear text structure ATS software expects. If you’re using a Canva template, check that it’s built single-column with selectable text you can highlight and copy before you trust it with an application.
What the AI screener reads once the parsing is right
Getting parsed correctly only gets you to the second filter. A language model scoring your resume checks each bullet for three things: an action, a number, and an outcome. “Managed a team” gives it none of the three. “Managed a 6-person team, cutting onboarding time from 3 weeks to 9 days” gives it all three in one line.
This is the X-Y-Z format: accomplished X, as measured by Y, by doing Z. It’s the single fix that matters most once your formatting is already clean, because a well-parsed resume full of vague duty statements still scores poorly.
Where to start
Start with the format, since a resume the parser can’t read never reaches the AI screener at all. Then rewrite your weakest three bullets in X-Y-Z format and see how much sharper they read. Our AI Screener Cheat Sheet walks through both in order, with a before/after rewrite bank. If your bullets are already drafted and you just want them scored and rewritten faster, the Resume Bullet Point Prompt Pack turns the same method into 8 copy-paste prompts. Or browse the full shop.