Big companies are building in-house AI design flows
Published 19 July 2026 · Source dated 15 August 2025
UX Collective rounded up how teams like Atlassian are wiring AI into specific steps of their design-to-prototype process, not replacing it wholesale. The pattern: narrow AI tasks get adopted, broad 'do my job' AI doesn't.
Atlassian's design team shared how they've built an AI-assisted design-to-prototype workflow, using AI at specific handoff points rather than one all-in-one tool. The pattern showing up across companies covered here: AI helps with speed at narrow tasks — renaming layers, drafting variants, summarising research — but humans still drive direction and judgement calls. This lines up with what other researchers have found: narrow, task-specific AI tools get adopted for real work; general-purpose 'AI does design for you' tools mostly don't. If you're trying to build your own AI workflow, look at where the repetitive, low-judgement steps live in your process (renaming, tidying, first-draft copy) and start experimenting there — not at the creative core of a project.
Coach
Map your last project's workflow and circle one repetitive step — try an AI tool there this week, not on the actual creative decisions.
AI now handles the boring bits of creative pipelines
Autodesk's SIGGRAPH 2024 tools show AI embedded in production pipelines, not just demos. Scheduling and rigging tasks get automated so artists spend more time on judgement calls.
adsknews.autodesk.com ·
Hype
AI wireframe copilots: useful for drafts, not decisions
AI design tool roundups this month show more studios using copilots for first-pass wireframes and filler copy. The tools still can't make UX decisions for you — that's still on you.
aidesign.tools ·
Hype
AI workflow tools are scaling — mostly outside design
A 2026 roundup of AI workflow automation platforms (Domo, ServiceNow, UiPath, Automation Anywhere) shows real enterprise adoption. But these tools are built for ops and data teams, not design-specific work.