Podcast Summary: The Ultimate Guide to AEO – How to Get ChatGPT to Recommend Your Product | Ethan Smith (Graphite)

Podcast: Lenny's Podcast: Product | Career | Growth
Host: Lenny Rachitsky
Guest: Ethan Smith, CEO of Graphite
Episode Date: September 14, 2025

Overview

This episode dives deep into Answer Engine Optimization (AEO), the emerging strategic discipline of getting your product, brand, or content recommended and cited by large language models (LLMs) like ChatGPT, Gemini, Claude, and others. Ethan Smith, one of the industry's top SEO/AEO minds, explains the evolution from classic SEO to this new world, offers tactical advice on how individuals and companies can get cited in generative answers, and breaks down the concrete steps—both on-site and offsite—for winning in this new channel.

Core Themes and Key Insights

1. The Rise and Definition of AEO (Answer Engine Optimization)

• AEO and GEO (Generative Engine Optimization) are essentially the same; both aim to have your product or content cited as answers by LLMs. “Answer Engine Optimization” is preferred for its clarity and focus (“answer” is more targeted than “generative”). (06:33–07:19)
• AEO is “SEO for ChatGPT”—the techniques overlap, but the mechanics and opportunities differ significantly.

Quote:

"AEO and GEO are essentially trying to describe the same thing, which is how do I show up in not LLMs as an answer? ...I personally prefer answer engine optimization."
—Ethan Smith (06:33)

2. Why AEO Matters Now

• LLMs like ChatGPT are rapidly becoming major traffic sources, in some cases surpassing traditional platforms (e.g., ChatGPT driving more traffic than Twitter for Lenny’s newsletter). (07:19–07:46)
• The transition is real, with substantial and speedy results possible—including for early stage companies.

Quote:

"You can get mentioned by a citation tomorrow and start showing up immediately... Early stage companies can win quickly."
—Ethan Smith (12:01)

3. Comparing AEO and SEO

• The head of AEO differs: It's about being cited most often across various sources (Reddit, YouTube, blogs) rather than just ranking #1. The most cited product wins in LLM answers. (10:37–11:47)
• The tail of AEO is much longer: People ask far more detailed, conversational, long-tail questions in chat than in search, creating opportunities for early, specific, and niche answers. (13:19–14:35)
• AEO can deliver significantly higher-quality leads. Example: "Webflow saw a 6x conversion rate difference between LLM traffic and Google search traffic." (14:42)

4. How to Win at AEO — Concrete Tactics

On-site Tactics

• Classical SEO Foundations Still Apply: Build high-quality landing pages that answer both the main query and anticipated follow-ups.
• Emphasize the Long Tail: Address specific, feature- or use-case-driven questions (e.g., detailed Help Center/support pages for tailored queries). (15:48–16:33, 60:53–62:55)

Offsite/Citation Tactics

• Show Up Everywhere LLMs Pull From: Get cited in video content (YouTube/Vimeo), Reddit threads, Quora answers, blogs, and especially in trusted affiliate and review sites (e.g., affiliates like Dotdash Meredith, Glamour, Good Housekeeping for B2C; TechRadar for B2B). (16:33–18:56, 38:44–38:57)
• Reddit is Key—But Can't Be Spammed: Authentic and transparent engagement beats attempted manipulation. (16:45–18:56)

Reddit Tactic Quote:

"Make a Reddit account, say who you are, say where you work, and give a useful answer. ...You don’t need 10,000 comments—even five could be great."
—Ethan Smith (18:56)

Video Tactic:
YouTube works especially well for B2B “boring but valuable” areas—there’s little content there, so filling that gap is high-impact. (29:12–29:52)

Tracking and Iteration

• Set Up Answer Tracking ("Share of Voice"): Like keyword rank tracking for SEO, but for LLM answers. Use trackers to see how often you're cited across multiple prompt variations and LLMs. (34:03–36:03)
• Experiment-Driven: Always run control/test groups, because LLM outputs are variable. Iterative, reproducible experiments are critical (test, measure, refine). (44:19–46:15)

Experimentation and Team Structure

• Divide questions into test/control, intervene with new strategies (YouTube, Reddit, affiliates) and measure statistical change in share of voice and traffic.
• Typical SEO teams can own most on-site strategy, but may need community/marketing specialists for offsite (content, Reddit, etc). (32:58–33:36)

5. Special Opportunities & Considerations

• Help Center Optimization: Tailor support content to cover niche, intent-specific questions—the “tail”—to win unique citations. Move to subdirectories, crosslink pages, and open up community Q&A as needed. (60:49–62:55)
• Blocking Model Training (But Allowing Indexing): If you don’t want LLMs to use your content as training data but want to be indexed/cited, manage through robots.txt and bot controls. (42:08–42:57)
• B2B vs. Commerce: Citation sources and tracking differ; for B2B, last-touch traffic is less useful, and post-conversion surveys are needed to attribute impact. (38:44–41:12)

Notable Quotes & Moments (w/ Timestamps)

On the Shift to AEO

• "Anything can be optimized. You just need to understand the underlying systems and rules of the game." —Ethan Smith (08:56)
• "For early-stage companies, my recommendation is don't do SEO at all. For answer engine optimization—definitely do AEO and only do citation optimization and long tail." —Ethan Smith (41:12)

On Tracking and Experimentation

• "Get your questions, turn tracking on, give it a couple weeks, make your changes, have a test group, have a control group. ...If the change worked and the control group didn’t, it worked." —Ethan Smith (44:19)

On AI-Generated Content

• "We found there's more AI-generated content on the internet than human-generated content... which is kind of disturbing." —Ethan Smith (51:59, 56:50)

On the Future

• "LLMs and search are going to converge...eventually you'll have LLMs remembering your context and making decisions for you, agents acting on your behalf." —Ethan Smith (58:55–60:25)

Mini-Step Plan for Winning at AEO (from Ethan's Recommendations)

1. Identify Key Questions

• Use your own and competitors' paid search data, convert to questions
• Use ChatGPT to help rephrase into common questions (29:12)

2. Set Up Tracking

• Employ answer tracking tools to monitor your citation rate/share of voice (36:03)

3. Analyze Current Citations

• See which sources and types of pages currently get cited

4. Landing Page Creation

• Produce or improve on-site content, especially long-tail and follow-up Qs

5. Offsite Citation Strategy

• Target YouTube, Reddit, Quora, tiered affiliates, and blogs; be authentic on communities

6. Design and Run Experiments

• Assign control/test, make changes, observe data, iterate based on results

7. Assemble the Right Team

• Blend classic SEO with community/content/offsite marketing expertise

Under-Discussed or Surprising Points

• Misinformation in the Space: Extreme ratio of hype to accurate info. “Google search is dying” is an evergreen myth—Google’s own traffic to publishers is steady. (47:18–50:43)
• Massive Overpricing for Commoditized Tooling: Many answer trackers charge enterprise prices for what is now a commodity, due to the complexity and novelty of AEO. (47:18–50:43)
• AEO ≈ SEO...with Key Differences: Most fundamentals are similar; the “head” (citation frequency) and “tail” (question diversity) are the meaningful changes. (End of episode)

Practical Nuggets

• Most SEO content is wasted: “19 out of 20 landing pages drive little to no traffic; 85% of impact comes from 5% of pages.” (25:19–27:23, 69:15)
• Optimization for LLMs is immediate: Unlike classic SEO, new companies can win on LLMs/answer engines the day after launch, simply by being cited. (12:01–13:19)
• AI-Assisted Content Works, Fully Automated Does Not: Human editing and originality are still key—all-AI spam gets identified and deprecated fast. (51:59–57:54)

Actionable Takeaways

• Start thinking of your SEO and AEO strategies as intertwined, but recognize and pursue new offsite and question-tail opportunities unique to AEO.
• Prioritize answer tracking and experimentation—don’t trust untested “best practices.”
• Engage authentically on channels LLMs trust: Reddit, YouTube, reputable blogs/affiliates.
• Expand Help Center/support content to win long-tail, intent-specific queries.
• Monitor traffic sources—you’ll see LLMs rising in importance, and the companies that figure out AEO first will reap compounding returns.

Additional Resources Recommended by Ethan

• Books:

  • Emotional Intelligence (Goleman)
  • Persuasion (Robert Cialdini)
  • How to Measure Anything (Douglas Hubbard)

• LinkedIn:

  • Ethan’s LinkedIn for original research and insights

• Graphite Blog (“5% Blog”):

  • https://graphite.io/5% – Dedicated to only the tactics that move the needle

Final Quote

“If you just say, don’t look at any of my data, then you cannot show up and your competitors will.”
—Ethan Smith (42:08)

Timestamps for Important Segments

| Timestamp | Segment / Highlight | |---------------|-------------------------------------------------------------------------------------------| | 06:33–07:19 | Definitions: AEO vs GEO | | 10:37–12:01 | How the “head” is different in LLMs vs Google | | 13:19–14:35 | The long tail: The explosion of new types of LLM questions | | 14:42–15:19 | Why LLM-driven leads are more qualified (Webflow 6x example) | | 16:45–18:56 | Reddit as a key offsite lever and how to engage | | 21:54–24:40 | Foundational principles: LLMs + RAG, topics, question research, citation optimization | | 29:12–34:03 | Step-by-step actionable plan for AEO | | 36:03–37:28 | Answer tracking: Share of voice and tools | | 38:44–41:12 | How strategies differ for B2B, commerce, and early-stage startups | | 42:08–42:57 | Should you allow LLMs to index/train on your content? | | 44:19–46:15 | Experimental methodology: control/test groups, reproducibility | | 47:18–50:43 | The most surprising/misinformation-filled aspects of AEO/AI | | 51:59–57:54 | The risks of AI-generated content and “model collapse” | | 60:49–62:55 | Help Center optimization: The power of answering long-tail, support-driven queries | | 63:19–68:19 | Lightning round (book, tool, and life mottos) | | 68:19–68:55 | Ethan’s proudest SEO/AEO win: Butter lettuce for Masterclass |

For more tactical AEO research and updates, connect with Ethan Smith on LinkedIn and check out the Graphite “5%” Blog.