Descriptions:
Dylan Davis presents a six-step AI workflow for building high-converting landing pages from research through deployment, with no coding required. The system is demonstrated using Claude’s Opus 4.5 model with extended thinking enabled, though Davis notes the same steps work with ChatGPT’s GPT 5.1 or Gemini 3 Pro with equivalent reasoning modes.
The workflow begins with a structured research prompt that instructs Claude to find current best practices for landing page conversion — with today’s date explicitly included so the model doesn’t surface outdated tactics from 2022 or 2023. Claude returns a markdown playbook covering layout structure, copywriting techniques, and visual design principles. In the copy phase, Claude uses that playbook alongside product details to write all page text. For visual inspiration, Davis captures screenshots from reference sites and uses Firecrawl to extract page content as structured JSON, giving the AI a concrete stylistic target.
With all four inputs assembled — playbook, copy, screenshot, and JSON — Claude generates a complete single-file HTML/CSS/JavaScript landing page rendered live in the artifacts canvas. For iterative refinements such as color adjustments and image placement, Davis recommends switching to a dedicated coding tool outside Claude to preserve the primary model’s context window. The video closes with a deployment walkthrough. It is aimed at solo founders, freelancers, and agency operators looking to eliminate developer costs while maintaining conversion-focused output quality.
📺 Source: Dylan Davis · Published December 13, 2025
🏷️ Format: Workflow Case Study
