Austin Lau had never opened a terminal. He had to Google how to do it. And for the better part of a year, he was the entire growth marketing operation — paid search, paid social, app stores, email, SEO, all of it — at a company that just aired two Super Bowl commercials, signed a $200 million partnership with Snowflake, and is currently valued at something north of $60 billion.
That fact alone is worth sitting with for a minute, because the instinct is to speed past it toward the tactical bits — the Figma plugin, the Google Ads CSV generator, the slash commands. And those are worth getting into. But the reason this matters to anyone reading a site called Future of PMM is that Lau's story isn't about one marketer's productivity hack. It's about the shape of the work changing underneath all of us, right now, whether we've noticed or not.
What Lau Actually Built
Anthropic published the case study in January, and the numbers are clean enough to summarize quickly. Lau used Claude Code — a command-line AI programming tool he had zero prior experience with — to build two workflows that collapsed the distance between having an idea and putting it in market.
Figma Plugin for Ad Variations
Built in ~45 minutes. Generates dozens of ad creative permutations across multiple aspect ratios with a single click. Replaced a manual copy-paste loop that used to eat 30 minutes per batch.
Google Ads Copy Workflow
Triggered by a custom /rsa slash command. Cross-references campaign data against brand tone and RSA best practices, exports 15 headlines + 4 descriptions per ad in an upload-ready CSV.
What used to take 30 minutes per ad now takes 30 seconds. Those are nice numbers. But the case study buries the more interesting data further down — the numbers from the rest of Anthropic's marketing org, which has since grown beyond Lau's solo run.
Why This Maps to What We're Building Here
If you've been following the content on this site — the agentic PMM piece, the Data Layer Wins positioning guide, Soo Shim's BattleCoach prototype — you've been watching us lay out a thesis in real time. The thesis is that AI doesn't just make product marketers faster at the stuff they already do. It changes what a product marketer is. The role evolves from someone who produces deliverables and manages dependencies into someone who identifies problems, builds solutions, and ships them — often without filing a single ticket.
Lau said it plainly in the case study: "I think growth marketing is going the way of almost like a product manager. We're not only able to execute on campaigns, we're able to actually build products in order to help us achieve our targets." That's not a metaphor. He literally built a Figma plugin and an automated ad pipeline. A marketer who can't code built software tools in a week.
This is the same pattern Soo demonstrated with BattleCoach — taking a workflow that depended on six handoffs and a prayer, and collapsing it into an intelligent system that one person can operate and iterate on. It's what we've been calling the 10x PMM: not ten times more decks, but ten times fewer dependencies between strategic insight and market-facing execution.
How to Think About Our Own Team
We run a 53-person product marketing organization that feeds content, positioning, and competitive intelligence into a demand gen engine. The workflow is familiar to anyone who's done this at scale: PMMs produce the narrative assets, demand gen distributes them, we measure what resonates, and we iterate. In between those steps lives a dependency chain — design builds the asset, web dev publishes the page, marketing ops configures the campaign, analytics reports the results. Every handoff introduces latency. Every latency point is where strategic intent starts to drift.
What Lau's story demonstrates — and what we're seeing early versions of on our own team — is that the dependency chain doesn't have to be a fact of life. It's an artifact of a world where the person who understood the message couldn't also build the vehicle for delivering it. That world is ending. Not because the specialists disappear, but because the person closest to the problem can now close the loop without waiting for the specialists to have bandwidth.
"A few years ago, if you had an idea to build something like this, you would probably need a team of engineers. Now, as a non-technical marketer, I can actually go out and build these things. The gap between 'I wish this existed' and 'I can actually build this myself' is much smaller than people realize."
— Austin Lau, Growth Marketing, AnthropicThe question for a team like ours isn't "how many people do we need" — that's always been the wrong question and it's especially wrong now. The right question is: what should 53 people be doing when the mechanical execution layer compresses by 5x? The answer, which we've been writing toward in every chapter of The Future of Product Marketing, is that they should be doing the things that actually matter — the strategic reasoning, the market sensing, the narrative construction, the buyer empathy work — and spending almost none of their time on the production mechanics that have historically consumed 60 or 70 percent of every PMM's week.
Think Like Lau
There are a few things about Lau's approach that are worth extracting as principles, because they apply to any PMM at any level — not just growth marketers with access to Claude Code.
🔑 The Lau Playbook
- Start with the friction, not the technology. Lau didn't set out to "use AI." He set out to stop wasting 30 minutes per ad on copy-paste loops. The tool was the means. The irritation was the catalyst. Every PMM has a version of this — the battlecard update that takes four hours, the campaign brief that requires six approvals, the competitive digest that nobody reads because it arrives three days late.
- Build something small first. His first Claude Code project was a calculator app. Not a marketing tool. A calculator. He needed to understand how the interaction model worked before he could trust it with real problems. That instinct — prototyping on something low-stakes before going after the workflow that matters — is how you avoid the twin traps of over-investment and premature dismissal.
- Talk to it like a colleague, not a search engine. "You don't need to know how to code. All you need to know is how to explain your challenge and what you're trying to solve in a very clear, concise manner." This is the skill. Not prompt engineering in the jargon-heavy sense — just the ability to articulate what's broken and what good looks like. PMMs should be naturals at this. We explain complex problems to non-technical audiences for a living.
- Bake in human judgment from the start. Lau was clear that all the copy and examples fed into his workflows were written by humans first — product marketing and copywriting teams. The AI brainstorms and iterates, but the baseline quality standard comes from people who understand the audience. This is the part that most "AI replaces marketers" takes miss entirely.
What This Means for the 10x PMM
We've been building toward a specific argument on this site and in the book — that the agentic era doesn't shrink the product marketing function, it elevates it. The PMMs who thrive will be the ones who stop defining their value by the volume of deliverables they produce and start defining it by the quality of the strategic decisions they make and the speed at which those decisions reach the market. Lau's story is the most vivid proof point we've seen yet that this isn't theoretical. It's already happening, at a company that has every incentive to demonstrate what's possible and every reason to be honest about how it works.
One person. Ten months. A $60 billion company. And the Figma plugin took 45 minutes.
If that doesn't change how you think about your own capability stack — what you invest your time learning, what you delegate versus build, how you define the value you bring to your team — read the case study one more time. Then open a terminal.
Read the Original Case Study
Anthropic's full write-up on how Austin Lau built his marketing workflows with Claude Code.
Read on claude.com →