Start with the business problem, not the redesign
Before touching tech, get clear on the operating capability you're actually trying to build. Driving demand, activating marketing, giving your team room to move — those are reasons to invest. "We haven't touched the site in seven years" isn't. Define the problem first.
Design systems: the foundation before the visuals
A brand deck isn't enough. Break your brand down atomically — typography, layouts, components — and today's tooling can understand it. Agents can generate variants that still feel like you, and your marketers can make the fine-tuned edits without waiting on a developer. Both need the same foundation to work from.
Content models: the least glamorous, most important work
Don't lift and shift 25,000 pages. Old content can quietly hurt your SEO and GEO performance. Audit first, rationalize, then build a content model around what survives. Structure your content — write once, reuse everywhere — and you give your team and your agents the context they need to keep it fresh.
Personalization: the bottleneck just moved
For 15 years, personalization stalled at scale. You had the strategy. You had the tech. You couldn't produce enough content. AI changes that. What used to feel like an impossible map of 60 audience variations is now something you can prompt to life and iterate on directly in the CMS.
AI delivery: pick your pill
AI makes projects faster. But enterprise customers carry decades of history, and that changes the equation. The real question isn't "how do we go faster?" It's "what capability do you want at the end?" Answer that first, then embed AI into it. The foundation shapes everything.
AI didn't make the foundational work optional. It made it the whole game.