<\/p>\n
Design organizations have adopted a remarkable number of tools in a short window \u2014 Figma, FigJam, DevMode, and now AI layers on top of all of it. The tools work. That’s not the problem.<\/p>\n
The problem is that nobody owns what happens next.<\/p>\n
There’s a pattern I keep seeing: leadership approves the budget, individuals start experimenting, and within 90 days the org has fragmented into a dozen parallel workflows that don’t talk to each other. One or two enthusiasts are burning out trying to carry everyone else toward the goal post. Everyone else is either quietly falling behind or quietly waiting to be shown the way.<\/p>\n
This isn’t an AI-specific failure. It’s the same challenge DesignOps has always addressed \u2014 just with a new surface area and a much shorter adoption curve compressing the timeline.<\/p>\n
Sustainable AI adoption happens across three layers:<\/p>\n
Most organizations invest heavily in Layer 1, lightly in Layer 2, and do nothing about Layer 3. Then they wonder why adoption stalls six months after launch.<\/p>\n
One of the earliest symptoms of missing infrastructure is vocabulary fragmentation. Teams develop local language for the same concepts. New hires have no map. Knowledge stays locked in individual heads \u2014 and leaves the org entirely when those people move on.<\/p>\n
Shared vocabulary is what makes it possible for a new hire to be productive in 15\u201330 days instead of six months. It’s not glamorous work. It doesn’t make the demo reel. But it’s the foundation everything else is built on, and it’s the cheapest, highest-leverage investment you can make before building any other system.<\/p>\n
AI tools introduce new questions that most design orgs haven’t formally answered yet: What’s confidential? What’s a quality standard for AI-assisted work? Who gets attribution? What are we not using these tools for, and why?<\/p>\n
Without explicit answers \u2014 in writing, owned by someone \u2014 teams will answer these questions differently. And inconsistency at that level compounds fast.<\/p>\n
I’ve started calling the accumulated risk of inadequate adoption practice AI Adoption Debt<\/strong>. Like technical debt, it’s not visible until it’s expensive. Unlike technical debt, most orgs don’t even have a name for it yet \u2014 which means they’re not tracking it, not measuring it, and not paying it down.<\/p>\n DesignOps is the function positioned to own this work. Not because it’s glamorous \u2014 it isn’t \u2014 but because DesignOps already owns the operational infrastructure of the design org. AI adoption is just the newest domain in that scope.<\/p>\n Right now, in most organizations: nobody. The enthusiasts do their best. The stragglers fall further behind. Leadership looks at tool adoption metrics and sees green.<\/p>\n What’s needed is an explicit owner \u2014 someone with design knowledge, organizational authority, and an operational orientation. Not the most excited person. The most equipped person.<\/p>\nWho owns this?<\/h3>\n
Three concrete starting points for org and design leaders<\/h3>\n