How Atlassian actually builds AI into design workflows
Published 19 July 2026 · Source dated 5 August 2025
UX Collective rounded up how companies like Atlassian use AI in their design-to-prototype process, based on direct sessions with their design team. It's a useful reality check against vaguer 'AI is changing everything' takes you'll see elsewhere.
Instead of theorising, this piece reports what Atlassian's design team actually does — using AI to speed up prototyping and testing cycles, not to replace the thinking that decides what to prototype in the first place. That distinction matters: the workflow change is in *speed of iteration*, not in *who makes the decisions*.
We're labelling this **adoption**: it's a named company's documented internal practice, not a speculative forecast. For a junior, the useful bit isn't 'use AI' in the abstract — it's picking one stage of your process (ideation, prototyping, or testing) and being deliberate about where AI actually saves time versus where it just adds noise.
Coach
on your next project, insert one AI-assisted step into a single stage of your process (not the whole thing) and note in your case study what changed and what didn't.
Usability heuristics are still the fastest self-review
Nielsen Norman Group's decades-old heuristic evaluation method remains the quickest sanity check before you spend money on user testing. It's unglamorous but still standard practice across the industry.
nngroup.com ·
Adoption
Redesigning workflows beats bolting on AI
A controlled field study found startups that redesigned their whole workflow around AI generated 90% more revenue than teams that just used AI to speed up individual tasks. Piecemeal AI use barely moves the needle by comparison.
uxtigers.com ·
Hype
AI sketching tools: useful ideation aid, not a shortcut
A design podcast recently dug into using AI for visualization and sketching in creative practice — the takeaway is blunt: AI can assist ideation, but only with careful curation from the designer.