Skip to content
ClaudeClaude

How Anthropic uses Claude in Marketing

There’s an opportunity developing for marketing teams that adopt tools like Claude Code to spend less time on repetitive execution and more time on the things that matter to their job. Austin Lau, growth marketer at Anthropic, shares how he went from never having opened a terminal to building Figma plugins and automated ad generation workflows, without writing a single line of code. Check out the full case study to learn more: www.claude.com/blog/how-anthropic-uses-claude-marketing Stay tuned for more stories in the "How Anthropic uses Claude" series.

Austin Lauguest
Jan 27, 20262mWatch on YouTube ↗

CHAPTERS

  1. From non-coder to curiosity: why Claude Code didn’t click at first

    Austin Lau opens by explaining he had never coded before and initially couldn’t understand what Claude Code was even for. He frames his work context in marketing channels that rely heavily on visual creative production.

    • No prior coding experience; basic setup felt unfamiliar (e.g., opening Terminal)
    • Initial confusion about the purpose of Claude Code
    • Marketing work spans paid social and App Store channels
    • Creative work often happens in Figma with design team support
  2. The real workflow pain: repetitive Figma edits across many specs

    He breaks down the manual process of duplicating frames in Figma and shuttling copy from a Google Doc. The repetition multiplies quickly across variations and aspect ratios, turning simple updates into a time sink.

    • Copy/paste frames in Figma to create variants
    • Manually move headlines from Google Docs into Figma
    • Back-and-forth workflow creates friction and errors
    • Workload balloons with 10+ variations across multiple aspect ratios
  3. Decision to automate: using Claude Code to build a Figma plugin

    Austin describes the pivot from manual repetition to automation by asking Claude Code to help create a Figma plugin. The goal: paste copy once and generate all creative variants automatically.

    • Identified a repetitive task suitable for automation
    • Used Claude Code to help build a Figma plugin
    • Objective: single paste action to update many variants
    • Plugin designed to scale across multiple variations quickly
  4. Plugin walkthrough setup: selecting a frame and launching the tool

    He switches into a demo of the plugin inside a Figma project file. The first steps are selecting the target frame and opening the custom plugin he built.

    • Demo begins inside a Figma project file
    • Starts with an example static image creative
    • Selects the frame to be used as the template
    • Opens the Claude-assisted custom plugin
  5. Template parsing: ‘Identify Template Components’ in action

    The plugin automatically detects editable regions within the design template. Austin points out it can find multiple components like headline, CTA, and code block—then he chooses to update only headlines.

    • Runs ‘Identify Template Components’ to parse the frame
    • Detected components include headline, CTA, and code block
    • User can choose which components to update
    • Focus for this use case: headlines only
  6. One paste to rule them all: importing headline lists into the plugin

    Austin copies a prepared list of headlines and pastes it into the plugin. This replaces repeated manual insertion with a single bulk input step.

    • Uses a pre-written list of headline variations
    • Copies the list once from another screen/document
    • Pastes all headlines into the plugin input
    • Sets up generation without repeated per-frame editing
  7. Generate Ad Variants: auto-creating frames for each copy version

    Upon clicking ‘Generate Ad Variants,’ the plugin creates distinct frames for every headline variation. The result is an instant set of ad versions, each separated and ready for review or export.

    • Triggers generation with ‘Generate Ad Variants’
    • Automatically produces multiple creative variations
    • Each variation appears in its own individual frame
    • Eliminates manual duplication and text replacement
  8. Impact on speed: cutting ~30 minutes down to ~30 seconds

    Austin quantifies the time savings from automating a previously manual, repetitive task. He emphasizes how the workflow improvement scales with more specs and variations.

    • Old process: ~30 minutes of repetitive steps
    • New process: ~30 seconds with the plugin
    • Time savings increase with more ad specs and variants
    • Automation reduces tedious, error-prone work
  9. Practical advice: start small and automate the repetitive parts first

    He closes with guidance for newcomers: find a small repetitive task and begin there. Coding experience isn’t required—clarity about the problem and desired outcome matters more.

    • Look for repetitive tasks that are good automation candidates
    • Start with a small, easy project to build momentum
    • You don’t need to know how to code to begin
    • Communicate the problem and goal clearly and concisely
  10. Mindset shift: the gap between ‘wish it existed’ and ‘I built it’ is smaller than you think

    Austin ends on a broader takeaway about empowerment: with tools like Claude Code, building custom workflow solutions is more accessible than most people assume. The key is recognizing solvable friction and taking the first step.

    • AI-assisted building makes custom tools accessible
    • Reframes feasibility: building may be closer than expected
    • Encourages experimentation and self-serve problem solving
    • Highlights the value of clear articulation over technical background

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.