Anthropic's Marketing Lead on 8 Practical Ways to Use AI
Austin Lau joined Anthropic in 2024 as the company's first U.S. growth marketing hire. In two years, he built performance marketing, SEO, lifecycle, and attribution functions largely solo - while Anthropic's revenue climbed from $150 million to $7 billion. His tools? The same Claude model his team markets to the public.
At the AirOps Next Conference in New York on May 13, Lau shared what he's learned about applying AI to marketing work. His advice cuts through hype: there's no magic playbook, and the technology changes faster than most teams can keep up.
Here's what he recommends:
Start with what Claude does better or faster
Before automating anything, ask two questions: What can Claude do better than your team? What can Claude do faster? The sweet spot is when the answer to both is yes.
The first step is encoding your current process end to end. "You want to be able to encode your entire process end to end," Lau said. "Then you can start to understand, okay, what are the areas that we can potentially automate?" Some tasks won't qualify - if a tool doesn't expose an API, you can't pull data programmatically.
Compare Claude's output to your own
Provide examples of what good looks like, then test Claude's work against what you'd produce manually. This fast, scrappy approach lets you evaluate quality before scaling.
"Here's Claude's output versus my output if I were doing it manually. Does this meet the bar or not?" Lau said. It's a practical way to set standards without lengthy trials.
Break complex tasks into smaller prompts
A common mistake is dumping an entire brief into one prompt. Lau did this early on with Claude Code and found it prevented him from checking work before the AI completed tasks.
Instead, break commands into digestible pieces. Start with a proof of concept to validate the approach, then build incrementally. This lets you catch problems early and adjust direction without wasting effort.
Buy tools instead of building them
Just because AI can generate code doesn't mean your team should replicate existing tools. Evaluate the ROI of building versus buying - and factor in maintenance costs.
"Once you build it, you need to think about all the tech debt," Lau said. "Somebody has to maintain it. Who's going to continue building on top of it?" Anthropic follows this principle itself: the company uses Salesforce and Gong like everyone else rather than building in-house alternatives.
Build custom tools only for truly niche problems
The exception: when your use case is precise, targeted, and rare - a problem so specific that no vendor will solve it.
"If nobody else has the same problem, nobody's going to build it and that means that I'm sort of out of luck," Lau said. "So in that case, I'll build something myself." For everything else, buy.
Encode tribal knowledge into reusable skills
Teams transfer knowledge through training and walking new hires through processes. Lau sees this as knowledge that should be codified and shareable instead.
"I'm not seeing enough non-technical teams actually encode their knowledge," he said, "so that it shifts from being this sort of tribal knowledge known by few to now being something where we can actually share this skill." Subject matter experts in reporting or campaign briefs should pass their knowledge on more easily to other teams.
Build reusable workflows, not one-off prompts
Without encoded workflows, every new conversation starts from zero. Scaling AI adoption means building repeatable skills, not re-prompting from scratch each time.
"You can actually truly leverage AI in a way that's going to be repeatable and scalable as opposed to trying to re-prompt every single time you're starting a new conversation," Lau said. "That's a pretty painful process."
Hire for craft and AI curiosity
Functional expertise remains required for marketing hires. But candidates should also be actively experimenting with AI, not just using it for basic copy edits or headlines.
"We still need somebody who lives, breathes, eats, sleeps whatever function we're hiring for," Lau said. "But some of the other things that we're also looking for now are, is this individual AI-pilled or even AI-curious?"
Look for people who excel at their craft and understand where AI fits into their workflow - whether to supplement weak areas or strengthen areas where they already excel.
For marketing teams starting to experiment with AI, consider exploring AI for Marketing or the AI Learning Path for Marketing Managers to build skills systematically.
Your membership also unlocks:
