The ultimate guide to AEO: How to get ChatGPT to recommend your product

SUMMARY

Ethan Smith, CEO of Graphite and SEO expert, shares tactical insights on Answer Engine Optimization (AEO) to boost visibility in AI tools like ChatGPT, highlighting its rapid wins for startups and superior conversion rates over traditional SEO.

STATEMENTS

Ethan Smith has 18 years in SEO, starting with programmatic and commerce SEO in 2007, marking Google's shift from spam-tolerant to algorithmically strict practices as the biggest prior change.
AEO and GEO both describe optimizing for AI-generated answers in LLMs, with AEO preferred for its focus on textual responses over broader generative outputs.
ChatGPT now drives more referral traffic to Lenny's newsletter than Twitter, illustrating AEO's emerging impact since January due to increased adoption and clickable answer features.
Traditional SEO tactics like high-volume keyword landing pages provide free AEO benefits, but AEO extends beyond by emphasizing multi-citation mentions over single top rankings.
In AEO, winning "best tool for X" queries requires appearing in multiple citations, unlike Google where the top blue link suffices.
Early-stage startups can succeed in AEO immediately via citations from Reddit threads, YouTube videos, or blog mentions, bypassing SEO's domain authority delays.
AEO's long tail is larger than SEO's, supporting more specific, conversation-driven questions never before searched, allowing wins on niche topics.
Webflow observed a 6x higher conversion rate from LLM traffic compared to Google search, attributing it to heightened user intent from conversational priming.
On-site AEO involves traditional SEO plus detailed follow-up content on features, integrations, and use cases; off-site focuses on citations from videos, UGC like Reddit/Quora, and affiliates.
Reddit is heavily cited in LLMs due to its authentic, community-moderated content, making spammy automation ineffective while genuine, disclosed contributions succeed.
LLMs use RAG (Retrieval-Augmented Generation) for post-training search and summarization, making citation optimization more controllable than influencing core training data.
Key AEO principles include treating content as topics covering hundreds of questions, prioritizing originality via information gain over typical rewrites.
Hyper-SEOed content plagues Google with low-ROI pages; AEO avoids this by valuing unique expertise and research in citations.
To rank in AEO, identify high-value questions from paid search data, convert to question variants using AI, and track via specialized tools.
Effective page types for AEO include listicles, category pages, and tools answering sub-questions; optimize by cross-linking and subdirectory structures.
Off-site strategies vary: pay affiliates like Forbes for mentions, create targeted videos for unglamorous B2B topics, and authentically engage on Reddit.
Experimentation is crucial in AEO, as many "best practices" lack analysis; use control/test groups on question sets to measure share of voice improvements.
B2B AEO relies on tracking non-clickable mentions and post-conversion surveys, differing from commerce's shoppable cards emphasizing schema and reviews.
Early-stage AEO skips mid-tier SEO, focusing solely on citations and long-tail specificity for quick visibility.
Blocking LLM training bots while allowing indexing preserves AEO opportunities without competitors dominating.
Misinformation abounds in AEO, like claims Google search is dying; instead, new channels like LLMs expand the overall discovery pie.
AEO tools often overprice commodity functions like tracking, similar to basic keyword tools that remain inexpensive.
AEO adoption spiked in January due to LLM growth and enhanced clickability, creating a steep, unusual growth curve.
Graphite's study found 10-12% of Google/ChatGPT content is AI-generated but ineffective; AI-assisted editing works, pure AI does not.
Internet now hosts more AI-generated than human content, risking model collapse from recursive derivatives.
Infinite AI loops in RAG converge answers to singular opinions, eroding wisdom-of-the-crowd diversity.
LLMs and search will converge into hybrid experiences with features like personalized agents and autonomous planning.
Help-center optimization targets AEO's long tail by moving to subdirectories, improving internal links, and crowdsourcing obscure use cases.

IDEAS

AEO democratizes visibility for startups by enabling instant citation-based wins, unlike SEO's years-long authority build-up.
LLMs favor frequency of mentions over rank position, turning citation diversity into a competitive edge for broad queries.
Conversational AI queries unlock unprecedented long-tail opportunities, surfacing novel questions impossible in traditional search.
Reddit's anti-spam community acts as a natural filter, preserving LLM trustworthiness and rewarding authentic engagement.
RAG's post-training mechanism shifts optimization from elusive model influence to actionable search retrieval tweaks.
Information gain as a content metric prioritizes novel insights, aligning AEO success with genuine value creation.
AEO traffic's 6x conversion boost stems from users arriving with refined intent after iterative questioning.
Subdomains hinder help-center AEO; subdirectories enhance crawlability and question coverage.
Early AEO ignores mid-funnel SEO, channeling efforts into off-site buzz for rapid market entry.
Share of voice in AEO demands multi-run, multi-surface tracking to capture probabilistic answer variance.
B2B AEO measurement bypasses direct traffic, relying on surveys to attribute indirect branded searches.
AI content proliferation risks "derivative collapse," where models echo homogenized views instead of diverse wisdom.
Citation ecosystems differ by vertical: TechRadar dominates B2B, while lifestyle sites rule commerce.
Experiment reproducibility counters SEO/AEO's volatile "best practices," ensuring scalable tactics.
Help centers become high-ROI AEO assets by addressing uncharted user queries from support logs.
LLM convergence with search could spawn autonomous agents handling personalized tasks without user input.
Misinformation hype around AI killing SEO mirrors past false alarms about social platforms.
Video citations like YouTube fill B2B gaps in unglamorous topics, turning content creation into leverage.
Control groups reveal natural LLM fluctuations, preventing misattribution in AEO tests.
Original research and expertise signals boost citations, evading rewrite cycles plaguing Google.
AEO's steep adoption curve post-January highlights clickability's role in traffic surges.
Affiliates like Dotdash Meredith exemplify scalable, high-trust citation networks.
Question mining from sales/support uncovers AEO tail goldmines overlooked by keyword tools.
Pure AI content fails detection-wise, with low false positives confirming its ranking penalty.
Zen-aggression duality in work—intense practice plus mindful presence—fuels deep expertise.

INSIGHTS

AEO amplifies SEO's foundations while introducing citation multiplicity and expanded long-tail dynamics for faster, intent-driven growth.
Authentic community engagement on platforms like Reddit inherently combats spam, ensuring LLMs prioritize trustworthy sources over manipulated ones.
Conversion superiority in AEO traffic arises from conversational priming, transforming passive searches into qualified, high-intent explorations.
RAG's separation from core training allows controllable, immediate optimizations, sidestepping the opacity of model retraining timelines.
Prioritizing information gain over typicality fosters original content that sustains rankings amid rising AI derivatives.
Early-stage AEO success hinges on off-site velocity, enabling startups to leapfrog authority barriers through viral mentions.
Experimentation with reproducible controls is essential to validate tactics, as untested "wisdom" often yields wasted efforts in evolving channels.
Help-center evolution into AEO hubs captures niche queries, turning support data into proactive visibility assets.
Blocking training while permitting indexing strikes a balance, securing distribution without ceding ground to rivals.
AI content's dominance threatens model diversity, converging outputs to bland consensus and eroding informational richness.
Vertical-specific citations demand tailored strategies, recognizing B2B's non-clickable nuance versus commerce's schema-driven clicks.
LLM-search convergence promises personalized, agentic experiences, redefining optimization around user context and autonomy.
Misinformation inflates AEO's disruption narrative, but evidence shows additive channel growth preserving search's core role.
Share-of-voice metrics across surfaces account for answer variability, providing a robust gauge of AEO efficacy.
Intentional, focused practice—rooted in deliberate effort—outpaces raw intelligence in mastering complex fields like AEO.

QUOTES

"ChatGPT is driving more traffic to my newsletter than Twitter."
"You can get mentioned by a citation tomorrow and start showing up immediately."
"Webflow saw a 6x conversion rate difference between LLM traffic and Google search traffic."
"The LM is summarizing many citations and so you need to get mentioned as many times as possible."
"Early stage companies can win. They can win quickly."
"Reddit is a community where it's real opinions from people authentic and it's heavily managed by the community."
"Anything can be optimized, but if you're spamming it, they'll see that and they'll have a whole team looking at that."
"The tail is larger in chat than in search... the average number of words... around 25 words where versus Google... around six words."
"AI-generated content doesn’t work."
"There's more AI generated content on the internet than human generated content."
"If you feed in derivatives of derivatives... you will basically take the wisdom of the crowd and that will shrink and you'll have a single opinion on everything."
"Google's slice of the pie stays the same. The pie gets bigger."
"Most best practices, most blog posts are not correct. So, how do you set up an experiment?"
"It's not your choice whether to play the game. You are playing the game whether you want to or not."
"The future of content is clearly AI assisted... but purely 100% AI generated does not work."
"Help center optimization and the long tail... people in chat are asking follow-up questions."
"I'm going to be the most intentional about my practice and I'm going to be as intense as I possibly can be about that practice."
"One out of 20 landing pages drive roughly 85% of all your traffic."

HABITS

Convert paid search keywords into question variants using AI prompts for comprehensive AEO targeting.
Engage authentically on Reddit by disclosing identity and providing useful insights without scaling to spam.
Mine sales calls and support tickets for long-tail questions to inform help-center content.
Set up control and test groups in AEO experiments, tracking share of voice over weeks for reproducibility.
Prioritize information gain in content by adding unique research or expertise not covered elsewhere.
Create targeted YouTube videos for niche B2B topics lacking existing coverage to build citations.
Use post-conversion surveys in B2B to attribute indirect traffic from non-clickable AEO mentions.
Block LLM training bots via robots.txt while allowing indexing for balanced participation.
Focus practice intentionally on high-impact tasks, blending aggressive execution with zen presence.
Publish original research studies on LinkedIn to boost engagement and establish authority.
Cross-link help-center articles extensively to enhance internal AEO relevance.
Experiment with affiliates for controllable mentions, budgeting for high-ROI vertical-specific sites.
Track answer variance across multiple LLM runs and surfaces to measure true visibility.

FACTS

AEO traffic converts 6x better than Google search, as evidenced by Webflow's data.
10-12% of content in Google and ChatGPT results is AI-generated, yet it underperforms human-edited material.
Internet now contains more AI-generated than human-generated content, based on Common Crawl analysis.
Google's search traffic to publishers is stable or slightly up, despite AI overview hype.
Perplexity's citations overlap 70% with Google, versus ChatGPT's 35%.
Subdomains rank worse than subdirectories in AEO contexts like help centers.
False positive rate for AI content detectors is around 8% on pre-ChatGPT material.
Webflow derives 8% of signups from LLM traffic, positioning AEO as a top channel.
19 out of 20 landing pages drive negligible traffic in SEO/AEO.
AEO adoption spiked dramatically in January due to LLM clickability improvements.
Model collapse occurs rapidly when training on recursive AI derivatives.
Reddit is the most cited non-site source in LLMs for its authenticity.

REFERENCES

Graphite (SEO agency and research papers at graphite.io)
Lenny's Newsletter (podcast and blog)
Orkes (enterprise platform for workflows)
Vanta (compliance automation tool)
Great Question (UX research platform)
Webflow (no-code website builder)
YouTube (video platform for citations)
Reddit (UGC community for authentic mentions)
Quora (Q&A site for user-generated content)
Dotdash Meredith (media conglomerate including Good Housekeeping, Allrecipes)
Forbes (affiliate for paid mentions)
Perplexity (AI search engine)
Gemini (Google's LLM)
ChatGPT (OpenAI's conversational AI)
Claude (Anthropic's LLM)
TechRadar (B2B tech review site)
Yelp (local business reviews)
Tripadvisor (travel marketplace)
TikTok (social search platform)
SurferSEO (AI content detector)
Common Crawl (web dataset)
Looker (analytics tool)
Otter (meeting transcription with Zapier)
BigQuery (data warehouse)
Emotional Intelligence (book on psychology research)
Influence: The Psychology of Persuasion by Robert Cialdini (persuasion frameworks)
Thinking, Fast and Slow by Daniel Kahneman (behavioral economics)
How to Measure Anything (measurement in uncertain contexts)
The Last Dance (Michael Jordan documentary)
Lance Armstrong documentaries (aggression-focused)
UFC (extreme sports viewing)
Free Solo and other Alex Honnold/Jimmy Chin climbing films (zen craftsmanship)
Outliers by Malcolm Gladwell (10,000-hour rule and focused practice)
AI Content Study by Graphite (white paper on AI effectiveness)
Model Collapse paper in Nature (recursive data training risks)
AEO Tools list (60 tracking options)
The Ultimate Guide to SEO (previous Ethan Smith podcast)
Inside ChatGPT (Nick Turley podcast)
Why ChatGPT Will Be the Next Big Growth Channel (Brian Balfour podcast)
An Inside Look at Deel’s Growth (Meltem Kuran Berkowitz podcast)

HOW TO APPLY

Analyze competitors' paid search keywords to identify high-value money terms relevant to your product.
Use ChatGPT or similar to transform those keywords into natural question variants people might ask LLMs.
Select 100-200 priority questions and input them into an AEO tracking tool to baseline your current share of voice.
Examine existing citations for top questions across LLMs, categorizing them into on-site, video, UGC, and affiliate groups.
Create or optimize landing pages matching successful formats like listicles or tools, ensuring coverage of all sub-questions and follow-ups.
Develop internal links and move help-center content to subdirectories for better crawlability and question relevance.
Produce targeted YouTube or Vimeo videos addressing underserved B2B or niche topics to generate video citations.
Engage on Reddit by creating a real account, disclosing your affiliation, and posting helpful responses in relevant threads.
Partner with affiliates like Forbes or Dotdash for paid mentions in high-trust publications tailored to your vertical.
Design AEO experiments: split questions into control (no changes) and test groups (e.g., add Reddit comments or videos).
Run interventions on test groups, wait 2-4 weeks, then compare share-of-voice lifts against controls for efficacy.
Reproduce successful experiments multiple times across question sets to confirm reproducibility and scale winners.
Integrate AEO into team workflows: assign SEO for on-site, community/marketing for off-site, and track via surveys for B2B attribution.

ONE-SENTENCE TAKEAWAY

Master AEO by prioritizing authentic citations and long-tail questions to rapidly boost high-conversion AI visibility for any stage company.

RECOMMENDATIONS

Focus AEO efforts on citation frequency rather than single rankings to dominate summary-based answers.
Leverage Reddit for genuine contributions, avoiding automation to align with community-trusted signals.
Target long-tail queries from support data to capture unique AEO opportunities without competition.
Implement RAG-focused optimizations like diverse off-site mentions for immediate, controllable impact.
Use AI-assisted content editing, not pure generation, to maintain originality and avoid detection penalties.
Experiment rigorously with control groups to validate tactics, discarding unproven online best practices.
Optimize help centers for subdirectories and cross-links to excel in feature-specific follow-up questions.
Track share of voice across multiple LLMs to account for answer variance and surface-specific differences.
Block training bots but allow indexing to participate in AEO without risking model dilution.
Prioritize video content for niche topics, filling gaps in B2B citation landscapes.
Mine sales and support interactions for question ideas, turning customer pain points into content wins.
Partner with vertical-specific affiliates early for scalable, high-trust mentions.
Measure B2B AEO via post-conversion surveys, capturing indirect traffic from branded follow-ups.
Embrace convergence by preparing for hybrid LLM-search experiences with personalized agent optimizations.
Publish original research to build expertise signals, enhancing citation quality over quantity.
Reproduce experiments multiple times to build a reliable AEO playbook tailored to your product.

MEMO

Ethan Smith, the CEO of Graphite and a veteran of 18 years in SEO, has witnessed seismic shifts in how we discover information online. From the spam-ridden early days of programmatic SEO to Google's algorithmic crackdowns, nothing rivals the current upheaval wrought by large language models (LLMs) like ChatGPT. In a recent podcast with Lenny Rachitsky, Smith demystifies Answer Engine Optimization (AEO), a strategy to surface products in AI-generated responses. Unlike traditional SEO, where domain authority builds slowly, AEO thrives on rapid citation accumulation—mentions across blogs, videos, and forums that LLMs summarize via Retrieval-Augmented Generation (RAG). This post-training search mechanism allows even fledgling startups to compete immediately, as a single Reddit thread or YouTube video can propel visibility overnight.

The allure of AEO lies in its potency: traffic from LLMs converts at six times the rate of Google searches, as seen in Webflow's 8% signup boost. Users arrive primed by conversational queries, narrowing intent through follow-ups on features and integrations. Yet, pitfalls abound. Smith's Graphite study reveals 10-12% of top results in Google and ChatGPT are AI-generated, but they flop—pure automation lacks the human touch that sustains rankings. Instead, authenticity rules: Reddit's community vigilance thwarts spam, favoring disclosed, helpful comments from real users. Off-site tactics shine here, from paying affiliates like Forbes for B2B endorsements to crafting niche videos that fill content voids in unglamorous sectors.

For practitioners, Smith's seven-step playbook starts with question mining—repurposing paid search data into LLM prompts via AI tools. Track performance with affordable AEO monitors, akin to keyword trackers, measuring "share of voice" across probabilistic answers. On-site, evolve landing pages into topic hubs covering sub-questions; off-site, diversify citations by vertical—TechRadar for SaaS, Glamour for e-commerce. Experiments are non-negotiable: split questions into control and test cohorts, intervene (e.g., Reddit posts), and replicate wins. B2B demands nuance, tracking non-clickable mentions through surveys, while commerce leverages schema for shoppable cards.

Misinformation muddies AEO's waters, with headlines falsely declaring SEO's death amid AI overviews. Smith counters: the discovery pie expands, Google's slice holds steady. Adoption surged in January, fueled by clickable modules, but tools peddle overpriced basics. Looking ahead, LLMs and search converge toward agentic futures—autonomous planners booking trips or recommending newsletters without prompts. Yet, recursive AI content risks "model collapse," homogenizing outputs into vanilla consensus.

Help centers emerge as AEO's sleeper hit, targeting long-tail queries from support logs. Shift to subdirectories, cross-link aggressively, and crowdsource obscure use cases via community forums. For early ventures, skip mid-tier SEO; chase citations and specificity. Smith's ethos—intentional practice blending aggression and zen—mirrors AEO's balance: hustle for mentions, but craft with originality. As LLMs hit billions of users, ignoring this channel cedes ground; embracing it, with experiments and authenticity, unlocks a high-ROI frontier.

In closing, AEO isn't revolution—it's evolution. Overlap with SEO persists, but citation velocity and intent amplification redefine wins. Companies like Deel scaled via Reddit pre-AI; now, it's essential. Block training bots to protect data, allow