In 2024, I defined the AI Marketing Technologist — a role to bridge marketing and AI. Eighteen months later, that framework is already too small. Not because the role doesn’t matter, but because the operating unit itself changed.

Economists are increasingly factoring in the possibility of rapid AI progress into their long-term forecasts on the job market.

A junior analysts can prompt their way to a beautiful Q1 deck in half the time I could five years ago. What they can't do — yet — is smell when the numbers are lying. That gap didn't exist when everyone had to wrestle the data themselves first.

Workslop refers to AI-generated work that seems polished but is flawed and in need of heavy corrections

AI struggles in enterprises not because of prompts, but because it lacks context. Here's how to build systems that fix that.

Everyone's "using AI" the same way everyone was "doing social media" in 2012 — a bunch of disconnected experiments with no one accountable for the output. Turns out deploying a chatbot and actually wiring AI into how your business makes decisions are two completely different projects.

I've been half-ignoring the "SEO is dead" takes for years because they're always wrong. But 30-40% conversion rates from LLM-referred traffic? That's not a trend piece. That's a revenue signal worth paying attention to — especially when most competitors aren't even tracking it yet.

People across Zapier have been scaling AI into every corner of how we work. Here are six common mistakes when scaling AI—and what to do instead.