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

AEO stands for answer engine optimization. It is the practical discipline of making sure your organization, product, or service shows up when people ask modern large language model tools—ChatGPT, Gemini, Claude, Perplexity—"what should I use?" or "how do I solve X?" For small businesses and nonprofits this is not a distant enterprise problem; it's an immediate opportunity. You can start showing up in answers fast, often faster than traditional search rankings allow, and the leads that come from these answer engines tend to be highly qualified.

Why AEO matters now (and why it is different, not completely new)

Search is changing but it is not dying. Think of AEO as LLM plus RAG: a large language model that summarizes a set of retrieved documents. The new answer experience is a summary built from citations—URLs, videos, forum posts, knowledge bases—and the model synthesizes them into a compact answer. That fusion is the thing you can influence.

Important high-level implications:

• It’s related to search. AEO uses the same basic signals search has always used: relevance, authority, and quality. But the mechanism for surfacing results is a summary rather than a single blue link.
• Early-stage orgs can win fast. You do not need years of domain authority to be cited in answers. A single well-placed mention on Reddit, a YouTube review, or a niche blog post can get pulled into an answer quickly.
• Quality of leads is often higher. When people ask conversational questions, they are primed and have more intent. One company saw a 6x conversion rate for LLM-driven signups versus Google search traffic.

Core concepts: head vs tail and citations

Two simple ideas let you think clearly about winners and losers in this new world.

Head vs tail

In classic search, controlling the top blue link for a head term generated disproportionate traffic. In LLM answers the head still matters, but differently. The model tends to mention whichever brands, tools, or sources appear most frequently across the citation set. If you are mentioned more often across citations you are more likely to be the chosen answer, even if you are not the #1 blue link.

The tail is much bigger with conversational systems. Whereas traditional search queries averaged a handful of words, chat prompts are long and include follow-ups. People ask highly specific questions that never made sense as a standalone Google query. That creates many niche opportunities for organizations that can answer the long tail well.

Citations are the new currency

Answers are synthesized from citations. Those citations come in many forms:

• On-site pages and help articles
• Video content (YouTube, Vimeo)
• User generated content (Reddit, Quora)
• Affiliate and publishing sites (Forbes, Dotdash Meredith properties)
• Documentation, product pages, integrations, and knowledge base content

Your job is to increase the number and quality of citations that mention your brand or product in a relevant way.

AEO for small businesses and small nonprofits: a practical lens

Small teams and nonprofits have constrained budgets and limited staff time. AEO is actually a good fit: you can use authenticity, direct community engagement, and well-targeted content to be cited quickly. You do not need huge budgets to start. Prioritize tactics that scale with effort rather than money.

Here’s how to think about priority:

• Immediate, low-cost wins: Reddit or niche community answers, YouTube videos you can produce in a few hours, help center articles that answer specific use cases.
• Pay-to-win, high-control plays: Affiliate mentions on trusted publishers, sponsored reviews, or paid placements on media properties.
• Foundational investments: Topic-driven landing pages and support documentation that compound over time.

The tactical AEO playbook (7 practical steps)

Below is a compact, repeatable playbook you can execute on a tight budget. It’s written for a small team or nonprofit with limited marketing hours per week.

1. Decide which questions you want to own

   Start with money terms and frequent user questions. Convert keywords and competitor paid search terms into conversational questions. Use whatever data you have—site search logs, support tickets, sales call transcripts. Feed lists into a generative model to expand each keyword into plausible question variants (but do the thinking: you know your customers best).

2. Track your share of voice across LLM surfaces

   Use an answer-tracking tool to measure how often your organization is showing up as a cited answer. The metric to watch is share of voice: what percent of runs of a question include you in the citations or in the final answer. Ask each question multiple times and on multiple surfaces (ChatGPT, Gemini, Perplexity, etc.) to measure variance.

   Pick the cheapest reputable tracker that supports the surfaces you care about. This will be your experiment baseline.

3. Audit who’s being cited today

   For each target question, collect the current citations and page formats that appear. Are answers pulling from listicles, official docs, YouTube videos, Reddit threads, or affiliate reviews? Reverse engineer the page type and the content patterns that win, then decide whether you can make a better version.

4. Create or optimize landing pages that answer the topic

   When you build a landing page, think topic not keyword. One page should aim to answer the range of follow-up questions people might ask. Structure your content so that each sub-question has a clear, scannable answer. Add short FAQs and example use cases.

5. Optimize off-site citations

   This is where smaller teams can punch above their weight. Target the specific citation types you found in step three and pursue them directly:

   • Make short, targeted YouTube videos that answer common questions; optimize titles and descriptions as conversational queries.
   • Engage in Reddit or niche community threads like an authentic member: identify yourself, state your role, and add a useful answer.
   • Work with affiliates or pay for a single authoritative mention if your budget allows.
   • Encourage customers to mention you in reviews and case studies where relevant.
6. Instrument experiments with control groups

   Take 100-200 target questions. Randomly assign half to be test targets (you will intervene) and the other half as control (do nothing). Track share of voice and average rank before and after interventions. Compare test vs control. Reproduce the experiment multiple times to validate what works.

7. Build cross-functional team ownership

   Answer engine optimization sits at the intersection of SEO, community, content, and product support. For small organizations this can be one or two people wearing multiple hats, but you still want clarity:

   • SEO lead: on-site pages, topics, tracking
   • Community marketing: Reddit, Quora, forums
   • Video/content specialist: short YouTube/Vimeo
   • Support/product: help center optimization and tail content

Hands-on tactics that actually move the needle

Not all tactics are equal. Small organizations should focus on these high-return activities first.

Help-center optimization: a surprisingly rich source of wins

Many chat queries are follow-ups—"does this product support X integration?" "Can it do Y use case?" Help center articles answer exactly those questions. Small nonprofits and businesses can quickly gain citations by doing three simple things:

• Move docs to a subdirectory rather than a subdomain where possible. Subdirectories share domain authority more reliably.
• Cross-link help articles logically so the topic clusters look authoritative.
• Mine support logs and sales questions to populate the long tail. Create pages or articles for real, specific scenarios customers ask about.

Community-sourced tails are powerful too. Invite users to submit their use-case articles and case studies so your knowledge base grows with the real-world, specific questions that chat will pull from.

Reddit: small authenticity beat mass spam

Reddit is heavily cited in LLM answers because of authentic user opinions. Resist the urge to scale fake accounts. That tends to get banned and is low ROI. Instead:

• Create a legitimate account, include your real name and role, and answer threads with useful guidance.
• Find threads that would be part of the citation set for your target question and provide value first. If appropriate, mention your product or service transparently.
• Five to ten thoughtful comments on the right threads can be more impactful than hundreds of shallow posts.

YouTube: a high-opportunity surface

There are fewer B2B videos answering very specific niche questions. Making short explainers or "how to" tutorials that directly address a question can put you into the citation graph. Videos are attractive to LLMs because they provide structured information and signals about what's being discussed in the description, captions, and metadata.

Tracking and measurement: how to know if you’re actually winning

Answer tracking differs from keyword rank tracking. Expect variance: the same question can produce different answers on each run, and different surfaces produce different citation sets. Your tracker should measure:

• Share of voice: the percent of runs in which you are included in citations or in the model’s final answer.
• Average rank across runs where rank represents how prominently you appear in the answer or citation list.
• Surface breakdown: which systems (ChatGPT, Gemini, Perplexity) and which page types (video, Reddit, docs) are sending traffic or citations.

To validate causality, use control groups as described earlier. Without a control you will mistake seasonal or adoption-driven lifts for the effect of your intervention.

Experiments that work and experiment design

Set a hypothesis, control and test groups, and a reproducible cadence:

1. Pick 200 target questions; randomly split into 100 control and 100 test questions.
2. Record share-of-voice and average rank for two weeks baseline.
3. Apply interventions on the test group only: publish a help article, post a video, or add authentic Reddit answers.
4. Monitor for 2–4 weeks and compare results between test and control groups.
5. Reproduce the test: perform the same playbook against a new batch of questions to confirm.

Reproducibility is the only guard against mistaken best practices. Many "rules" circulating online are one-off observations that do not hold up under systematic testing.

AI-generated content: proceed with caution

There are two realities that are relevant to small teams and nonprofits.

• AI-assisted content is useful. Use generative models to draft, iterate, and speed up writers. Human editing, domain expertise, and original research remain essential.
• Fully automated AI content performs poorly in the long run. Studies show that pure AI-generated content does not reliably rank or sustain performance. Search and answer systems are designed to favor human-sourced, high-information-gain content as they adjust to avoid spam and derivative loops.

Why this matters: if every player publishes derivative machine-generated pages en masse, the models will start to converge on recycled opinions and lower information gain. That outcome hurts everyone, and platforms will modify algorithms to deprioritize derivative content. The sustainable path is human-in-the-loop content with original information and perspective.

Attribution pitfalls and how to measure real impact

LLM answers can be non-clickable for B2B contexts. When a tool summarizes and mentions you, the user might open a new tab and perform a branded search or go direct. Standard last-touch referral metrics will misrepresent the impact of AEO. For reliable attribution:

• Track share-of-voice in answer trackers rather than relying solely on referral logs.
• Ask new customers “how did you first hear about us?” during onboarding or sign-up flows and record answers.
• Use experiments with control groups to infer lift rather than only measuring raw referral counts.

How strategy differs by business type

Not all sectors should use the same tactics.

B2B SaaS

Answers for B2B tend to cite product reviews, specialized publications, and technical docs. Many answers are not clickable. Optimize help centers, integration pages, and produce detailed technical explainers. Track long funnels—B2B buyers often require tens of touchpoints before conversion.

Commerce and marketplaces

Commerce has more clickable modules and shoppable cards. Schema, reviews, product feeds, and merchant data are critical. If you sell product directly, invest in rich structured data and make sure marketplaces or retailers listing your products are properly configured.

Early-stage organizations and nonprofits

Early-stage orgs should prioritize citation optimization and tail content. Create a handful of extremely focused content pieces and community answers that show up for the obscure but high-intent questions your audience asks. You can win fast here because you do not need years of domain authority to be cited.

Team and tooling: who should do what

For a small team, roles overlap, but you want coverage across these skill areas:

• Topic owner (SEO/content) — builds landing pages, organizes topics and FAQ content.
• Community marketer — authentic engagement on Reddit, Quora, niche forums.
• Video/content creator — short form answer videos targeted to niche queries.
• Support/product owner — curates help center content and mines support questions.
• Analyst — runs experiments, tracks answer share of voice, and measures lift.

Start with one generalist who owns the playbook and expand as you see returns.

Quick win checklist for small organizations

• Pick 10 high-priority questions and map where answers currently come from.
• Create or revise a help center article for one of these questions and move it to a subdirectory.
• Make one short YouTube video that answers a specific long-tail question and optimize the title as a question.
• Post 3 authentic Reddit answers where your audience congregates; identify yourself clearly.
• Track the 10 questions in an answer tracker weekly for changes in share of voice.

Long-term risks and the future

There are structural risks as LLMs and search converge. Model collapse—where models trained on derivative content start to degrade—is a real research concern. Platforms will adapt to avoid derivative spam, likely increasing the weight of original research, authoritative voices, and data-based signals.

Expect convergence between search and LLM-driven answers. Google, Bing, and LLM platforms will borrow features from each other. That means the fundamentals—originality, helpfulness, and trust—remain the durable levers you can control.

Case example: what success looks like

A concrete example: a design platform saw an 8% share of signups from LLM-driven traffic. In another instance, an organization observed a 6x higher conversion rate for LLM referrals compared to classic search. The consistent theme in both cases was topic-driven content on site plus a diverse citation footprint across videos, forums, and authoritative mentions.

Common mistakes to avoid

• Assuming AEO is completely different from SEO. It is not. Many proven SEO principles carry over.
• Trying to spam Reddit or buy fake engagement. Community moderation catches it and the value is low.
• Relying on fully automated AI content. Human oversight and original information remain essential.
• Failing to instrument experiments and treat anecdotes as facts. Run reproducible tests.

Priority roadmap for the next 90 days (for small businesses and nonprofits)

1. Week 1–2: Audit 20 top user questions; set up tracking for 20 targets.
2. Week 3–4: Publish or update 5 help center articles and 2 long-form landing pages targeting topics.
3. Week 5–8: Create 3 short YouTube videos and seed 10 authentic community answers on Reddit or niche forums.
4. Week 9–12: Review tracking, compare test vs control, iterate on top-performing tactics and scale what works.

FAQ

How is answer engine optimization different from traditional SEO?

AEO is not the opposite of SEO. It builds on SEO fundamentals but focuses on being cited inside LLM answers. The important differences are that LLM answers synthesize multiple citations, the tail of questions is much larger due to conversational follow-ups, and off-site mentions (videos, community posts) have an outsized role in being selected as sources.

Can small nonprofits realistically compete in AEO?

Yes. Small nonprofits can win quickly by targeting very specific, high-value questions and being authentic in community spaces. Help center articles, testimonials, niche blog posts, and community engagement are low-cost ways to generate citations that LLMs will pull from.

Does letting LLMs index my content hurt me?

Indexation can increase discoverability and citations, which helps you show up in answers. If you are worried about being used for model training, many platforms provide ways to allow indexing but block training with specific user-agent or robots directives. You can choose to allow indexation but restrict training if that aligns with your strategy.

Are AI-generated landing pages a good strategy?

AI-assisted content can speed up production but fully automated AI-generated content without human editing tends to perform poorly and risks being deprioritized by platforms. Use AI to draft and accelerate, but invest human time into editing, adding original research, and improving information gain.

How should I measure impact if answers are not clickable for my product?

Use answer tracking for share of voice and average rank. Complement that with post-conversion surveys asking users how they first heard about you. Use tests with control groups to measure lift rather than relying only on last-touch referral metrics.

What is the quickest action that produces the best ROI?

For most small teams the best immediate leverage is help center optimization for specific follow-up questions and authentic community engagement in places like Reddit. Both are low cost, can be implemented quickly, and show up in answers faster than trying to build domain authority.

Which LLMs should I optimize for?

Optimize for the surfaces your audience uses. ChatGPT is large and growing fast, but other platforms like Perplexity and Gemini may have different citation patterns. Start with one or two surfaces that matter most and expand. Track across multiple systems to understand variance.

Final guidance: start small, iterate fast, be authentic

Answer engine optimization is a new channel built on old principles: be helpful, be original, and be discoverable. For small businesses and nonprofits this is good news. You can compete by being authentic in communities, by answering the real long-tail questions your users ask, and by investing a little time to instrument experiments. The returns are often meaningful—better-qualified leads and the chance to be the trusted recommendation someone hears when they ask "what should I use?"

Begin with a short experiment: pick ten questions, track them, publish one help article, make one short answer video, leave three authentic community answers, and measure share of voice. If you see lift, double down. The ability to show up in answers will get more important over time. Build the muscle now and your organization will be that much more discoverable at the moment people are actively deciding.

This article was created based on the video The ultimate guide to AEO: How to get ChatGPT to recommend your product | Ethan Smith (Graphite).