Building design
infrastructure for
the AI era

The situation

Eight years as Inman’s VP of Design, working directly with CEOs Josh Albertson and Emily Paquette and reporting to CPO Eliot Shepard. Inman is a mature, profitable media company — budgets are real, headcount decisions are deliberate, and there’s no venture capital to throw at problems. The mandate was to run creative at a level that punched far above what the budget could buy.

That’s a different constraint than I’d worked under before. At Vox Media, we had venture dollars — the answer to a capacity problem was usually to hire. At Inman, the answer had to be smarter. So when the pace of the organization outgrew a one-person department, I stopped trying to keep up and started building infrastructure instead — the kind that scales without adding headcount.

The frame

Design can’t be decoupled from business. Business can’t be decoupled from efficiency. Efficiency can’t be decoupled from AI. The question isn’t whether to adapt — it’s whether you build the infrastructure before or after the chaos arrives.

At Inman, I built it during the chaos. While the engineering team was rebuilding how they worked with AI, design couldn’t afford to be left behind — or worse, become a bottleneck in a workflow that was moving faster every week. The only way to stay relevant was to build infrastructure that could keep pace with how the whole organization was changing.

What I built

After Effects still renders the video. Illustrator still builds the badge. Photoshop still comps the ad. The craft tools didn’t go anywhere, what changed was everything around them. The brief, the brand constraints, the review process, the production handoff. AI compressed the distance between idea and execution. The hands doing the work are still human.

• 01 Brand standards site — The infrastructure layer built on top of Inman’s brand system. Hosted on GitHub Pages via Claude Code — always current, always accessible to people and agents alike. Paired with machine-readable brand documentation: design.md files that encode every design decision in a format both humans and AI can reference directly. The brand story lives in case 02. This is what happens after the brand exists.
• 02 Production workflow — Integrated thebrief.ai as Inman’s primary design production tool, cutting production time by 2/3 across event graphics, social, ads, and presentations. One person, no longer the bottleneck.
• 03 Creative brief governance — Documented the AI workflow process so the entire team could operate within it, not just me. Infrastructure only works if others can use it.
• 04 Event design system — Full design system for Inman Connect San Diego 2026. Room color environments, badge artwork, stage graphics, LED wall backdrops, interactive pattern explorer. Built with AI assistance throughout.
• 05 Tool evaluation — Currently testing Leonardo.ai to replace Visual Electric. The jury’s still out. This is the job now, continuous evaluation, rapid integration, no attachment to any single tool.
• 06 Machine-readable brand docs — The brand standards site is only half the infrastructure. The other half is design.md — a structured file that teaches Claude the entire Inman brand system. Not just values, but judgment calls: what’s universal vs. rare, what combinations are approved, how components behave. When AI generates anything for Inman, it works within documented constraints rather than guessing. This is what makes the system scalable without adding headcount.

The bigger point

This isn’t a story about AI making design easier. It’s a story about what becomes possible when a creative department stops operating like a service bureau and starts operating like an infrastructure team.

The tools will keep changing. Visual Electric is gone. Something else will be gone next year. What doesn’t change is the need for systems, documentation, and someone who knows how to build them fast and rebuild them faster.