This video newsletter by Christopher Penn discusses optimizing marketing for AI, emphasizing the measurability of referral traffic and sources used by generative AI tools. It covers technical optimization, including website accessibility, llm.txt implementation, and knowledge blocks. Content optimization focuses on multimodal content creation, answering user questions, structural elements, subtitles, captions, jargon, citations, and content freshness. Offsite optimization involves press releases, guest appearances, social media presence, reviews, and multilingual content. The newsletter concludes with the urgency of influencing AI models before synthetic data takes over, urging marketers to adapt their strategies for the AI-first era.
Metadata
- Type of Content: Newsletter Transcript
- Domain: www.youtube.com
- Organization: Trust Insights
- Date Published: March 2, 2025
- URL: https://www.youtube.com/watch?v=ThCCmTdwv8A
Summary
- Mythbusting: You can't determine brand placement or awareness within an AI model. AI tools aren't search engines; users interact with them conversationally, making brand presence measurement incredibly complex and impractical.
- What's Measurable: Referral traffic from AI tools to websites, and roughly, the sources AI tools use (due to search grounding). You can also track how often AI crawlers access your content.
- Technical AI Optimization: Optimize your website for screen readers and text browsers for better AI crawling. Implement llm.txt (like robots.txt, but for LLMs). Use International Phonetic Alphabet (IPA) for names and key terms. On YouTube, allow third-party AI training in your channel settings. Add knowledge blocks to your website content. Don't block AI crawlers in robots.txt.
- Content Optimization: Create content in multiple formats (video, audio, text, images, etc.). Anticipate and answer user questions within your content. Use structured HTML elements (H1, H2, lists) over just CSS styling. Use alt text for images and captions/subtitles for video/audio. Use industry jargon. Cite relevant experts and sources in your field. Keep content fresh, accurate, and prune outdated content.
- Offsite Optimization: Issue press releases for wide distribution. Appear as a guest on podcasts, YouTube channels, and other rich media platforms. Be active on social media platforms used for AI training data (Facebook, Instagram, Threads, LinkedIn, X/Twitter). Encourage reviews. Use deep research tools to identify key sources and platforms. Create content in multiple languages, especially Mandarin Chinese.
- Urgency: The window to influence AI models is closing as model makers shift to synthetic training data, which doesn't include your original content.
What makes this novel or interesting
- Emphasis on Conversational AI: Acknowledges how users interact with AI differently than search engines, shifting optimization focus.
- Focus on Multimodal Content: Stresses creating content in various formats for broader reach and training of multimodal AI.
- Introduction of llm.txt: Presents a new technical element for AI optimization.
- Strategic Use of Jargon and Citations: Advocates using jargon and citing experts to improve AI's understanding and association-building.
- Emphasis on Speed and Urgency: Highlights the limited window of opportunity to influence AI models before the widespread adoption of synthetic training data.
- Use of Deep Research Tools: Suggests using these tools to identify key information sources and optimize content placement accordingly.
Verbatim Quotes
What Not To Do
- "First and foremost, there is absolutely no way whatsoever to determine brand placement or brand awareness in an AI model—none, zero, zilch. Anyone claiming otherwise is either unaware of how AI works (. . .) or they're lying."
- "AI tools—generative AI tools—are not search engines. People don't use them like search engines."
What is Measurable
- "As the old adage goes, 'If you can't measure it, you can't manage it.' Even in AI, that is still largely true."
- "we can measure referral traffic from generative AI tools to our websites."
- "we can at least roughly measure what sources generative AI tools are using (. . .) because more and more tools are using search as a grounding function for AI.”
Technical AI Optimization
- "The number one thing you can do (. . .) with your website is to make sure it is well-optimized for AI by making sure it is well-optimized for anyone using a screen reader or other visual assistance tool."
- "For the technical optimization of your site, you're going to want to implement llm.txt, which is Anthropic’s LLM summary of your site."
Content Optimization
- "Today's content can't just be in one format. Multimodal AI models are training on everything they can get their hands on: Video, audio, images, and text."
- "When you create content, put it through the generative AI tool of your choice (. . .) to ask questions of the content (. . .) The goal is to determine what else should be in your content that users are likely to ask as a follow-up question in the AI tool of your choice."
Offsite Optimization
- "The memo here is dead simple: Be in as many places as you can be, however much time you can commit to being everywhere."
- "You want to be a guest on as many other people’s channels as possible (. . .) When it comes to distribution, prioritize rich media: Podcasts, YouTube channels, streamers—anything with video."
Wrapping Up
- "The window to significantly influence AI models is closing (. . .) because model makers have run out of content that they can use (. . .) they’re creating and feeding synthetic data (. . .) instead."
How to report this in the news
In a news broadcast, explain that the way we optimize content for discoverability is changing due to AI. It's no longer just about keywords for search engines. Now, we need to think about how AI "learns" about brands. Imagine training a dog: You wouldn't just tell it once to sit. You repeat the command, show it what you mean, and reward it when it gets it right. Similarly, we need to "train" AI by providing consistent, high-quality content in multiple formats, in multiple places, answering common questions, and associating ourselves with relevant experts. This report should emphasize the urgency of adapting to these changes, as the window of opportunity to influence AI is closing.
Detailed Recap (for marketers)
Website Optimization Checklist
- Accessibility is Key: Ensure your website is easily navigable and readable using screen readers and text-only browsers like w3m or Links.
- Implement llm.txt: Create and upload an llm.txt file to summarize your website for LLMs.
- Use IPA: Incorporate International Phonetic Alphabet notation for brand names, key executives, products, and services.
- Knowledge Blocks: Add concise, informative knowledge blocks about your organization within your website's main content template.
- Traditional SEO: Don't neglect basic SEO practices like schema.org markup and JSON-LD.
Content Strategy Checklist
- Multimodal Content Creation: Create content in various formats (video, audio, text, image, etc.) for maximum reach. Use the "video first" approach and repurpose video content into other formats.
- Answer the Questions: Use AI to identify potential follow-up questions users might have about your content and address them proactively.
- Structural Elements: Use proper HTML structural tags (H1, H2, lists) to define the structure of your content, rather than relying solely on CSS styling.
- Subtitles and Captions: Provide alt text for all images and captions/subtitles for audio and video content.
- Jargon and Citations: Use industry-specific jargon and cite relevant experts in your content.
- Content Freshness: Keep content up-to-date, accurate, and prune or retire outdated or irrelevant content.
Offsite Promotion Checklist
- Press Releases: Issue press releases on reputable wire services for broad distribution.
- Guest Appearances: Seek out guest appearances on podcasts, YouTube channels, and other rich media platforms.
- Social Media: Maintain an active presence on platforms like Facebook, Instagram, Threads, LinkedIn, and X (Twitter).
- Reviews: Encourage and solicit reviews on relevant review platforms.
- Deep Research Tools: Use deep research tools to identify where AI is getting its information and prioritize content placement accordingly.
- Multilingual Content: Create content in multiple languages to reach a wider audience and enhance AI training data, especially Mandarin Chinese.
Key Takeaways
- Urgency: The window to influence AI models is closing; act now.
- AI is Different: Treat AI differently than search engines. Focus on conversational AI and user experience.
- Be Everywhere: Maximize your online presence across multiple platforms and in multiple formats.
- Focus on Machines: Optimize content for AI consumption, not just human readability.
- Don't Neglect Traditional SEO: Search grounding is still important for many AI models.