Descriptions:
Craig Hewitt, founder of podcast hosting platform Castos, explains why strong Google SEO rankings do not translate to visibility in ChatGPT, Perplexity, Claude, or Grok — and walks through the seven concrete changes his team made to close that gap. Castos maintains over 400 pages of targeted content and ranks on Google’s first page for dozens of relevant keywords, yet was largely absent from LLM-generated answers, prompting a full content strategy overhaul.
The centerpiece of the video is SEO Machine, a free, open-source Claude Code project Hewitt has been developing for six to eight months. The tool automates LLM-optimized article creation by seeding Claude with detailed context about the business, its customers, and its unique value proposition. Key techniques demonstrated include leading every article with a direct answer in the first one or two sentences before any narrative framing, inserting a mandatory TLDR block immediately after the introduction, placing author authority signals above the fold rather than at the page footer, and running an AI citation research command that generates roughly 100 representative queries a potential customer might type into an LLM — then auditing which currently surface the brand.
The SEO Machine repository is publicly available on GitHub and designed to be cloned and configured with minimal setup, making the workflow accessible to content teams without dedicated engineering resources. Hewitt also discusses how authority signals and context seeding differentiate LLM-targeted content from generic AI-written blog posts.
📺 Source: Craig Hewitt · Published April 16, 2026
🏷️ Format: Workflow Case Study
