Question 1

Why does it now matter how ChatGPT and Perplexity describe my startup?

Accepted Answer

AI assistants are now where many buyers first ask "what should I use for X": half of B2B software buyers start their research in a chatbot, and zero-click search means fewer people ever reach your site. What ChatGPT, Gemini or Perplexity says about you decides whether you make the shortlist before you know the buyer exists. The clicks themselves are still few (AI platforms sent about 0.32% of all website traffic in 2026), so the influence sits in the answer, not in your analytics. If they describe you wrongly, or not at all, you lose deals invisibly.

Question 2

How do founders get their product recommended by AI assistants?

Accepted Answer

There is a repeatable playbook: pick the buying prompts you must win, publish the most specific content on those exact questions, and make sure the sources AI trusts (Reddit, review sites, listicles, YouTube) talk about you. Results can appear in weeks because assistants re-retrieve the web constantly, so early-stage startups can compete. Start with ChatGPT, which sends about three quarters of AI referral traffic, then Gemini, now second, and track your share of voice against competitors. ChatGPT now sells ads, kept apart from its answers, so recommendations are still earned.

Question 3

Is traditional SEO dead, what should an early-stage startup actually do?

Accepted Answer

No, but it has mutated. Google still sends far more visitors than all AI assistants combined (about 134 times more in one 2026 clickstream study), while AI answers eat top-of-funnel clicks (up to a 58% CTR drop when an AI Overview appears). Google says its AI features use its normal index, and its own guide, updated in July 2026, tells you to skip AEO hacks like chunking pages or adding llms.txt files. An early-stage startup should run lean, bottom-of-funnel SEO and layer AEO on top, rather than picking one and abandoning the other.

Question 4

How do I find out what ChatGPT, Perplexity, and Gemini currently say about my startup, and track it over time?

Accepted Answer

Start free: run 10 to 20 real buyer prompts (not just your own name) across ChatGPT, Gemini and Perplexity every month, logging mentions, position, sentiment and cited sources. Search Console's generative AI report, open to every site since 31 August 2026, shows how often your pages appear in AI Overviews and AI Mode. HubSpot's AEO Grader or Ahrefs' free checker give a quick baseline, and paid trackers start around $29 a month. What matters is visibility rate and share of voice against rivals, not one answer.

Question 5

How do AI assistants actually choose which sources to cite in an answer?

Accepted Answer

Two separate systems are at work: the model's training data (slow to change, favours brands widely written about) and live retrieval, where the assistant searches the web and cites pages it can read cleanly (ChatGPT runs its own OAI-SearchBot crawler and sometimes queries partner search providers). Citation studies show each engine has its own diet: Reddit and Wikipedia are heavy in ChatGPT, Perplexity rewards fresh structured pages, and Google's AI features pull from its normal index. The founding GEO research found that adding statistics, quotations and clear structure lifts your odds of being cited by 30 to 40%.

Question 6

Do Reddit, review sites, and press mentions matter more than my own website for AI visibility?

Accepted Answer

Often yes: the bulk of brand mentions in AI answers are grounded in third-party sources, Reddit is among the most-cited domains in ChatGPT and Perplexity, and G2 is the single most-cited platform for software queries. That is why the modern playbook is as much digital PR, review depth, and honest community participation as it is on-site content. Your own site still matters for facts about you, but assistants trust what others say about you more than what you say about yourself.

Question 7

Should I publish comparison pages and 'best X' listicles to win AI recommendations?

Accepted Answer

Yes, this is currently the highest-leverage content format: listicles and comparison articles take roughly a third of all AI citations, and around 40% on commercial 'best X' prompts. Write genuinely useful ranked lists that include competitors, put a direct answer in the first paragraph (models lift most citations from the top of a page), and keep them updated. Honest comparisons on your own domain plus placements in third-party lists compound each other.

Question 8

What technical setup does my site need so AI crawlers can find, read, and cite it?

Accepted Answer

Do not block the bots that matter: OAI-SearchBot decides whether you appear in ChatGPT search (GPTBot is for training, a separate choice), plus PerplexityBot and Google's crawlers. Serve server-rendered HTML, not JavaScript-only pages, and get indexed in Bing too. Agents like ChatGPT Work, Gemini's auto browse and Claude in Chrome now visit sites to finish tasks, so use clear labels, real buttons and accessible forms. Schema helps machines parse your facts, but Google says no special AI markup is needed and it does not use llms.txt, so that file is optional.

Question 9

Google now answers with AI Overviews and AI Mode. Does my Google strategy change?

Accepted Answer

Google says AI Overviews and AI Mode use the same index and ranking systems as normal search, so the entry ticket is unchanged: crawlable, helpful, people-first content with no special AI markup. The payoff has changed: informational clicks have collapsed (up to a 58% CTR drop when an Overview shows, and in an August 2026 experiment AI Mode cut clicks to outside sites by 18.8 points). In October 2026 Google added more inline links, and Search Console now reports your AI impressions. Aim to be the cited source, and put effort into bottom-of-funnel pages and content AI cannot paraphrase away.

Question 10

How do I measure traffic, signups, and revenue coming from AI search?

Accepted Answer

GA4 now files many AI visits under a built-in AI Assistant channel, but its list leaves sources out (Perplexity lands in Referral) and clicks from AI Overviews and AI Mode stay in organic search, so add a custom channel group matching referrers like chatgpt.com, perplexity.ai and gemini.google.com. Compare conversion rates rather than raw sessions: the volume is small but converts better, from 31% higher for ecommerce to 4 to 5 times in some studies. A large share of AI-influenced visits arrives with no referrer and lands in Direct, so whatever you measure is a floor.

Question 11

I sell products online. How do I get into ChatGPT Shopping, Perplexity, and agent-driven commerce?

Accepted Answer

OpenAI, Google and Perplexity all take product data directly. In ChatGPT, Shopify and Etsy stores are already included and other merchants apply to share a feed; OpenAI dropped in-chat Instant Checkout in 2026, so shoppers find products there and buy on your site, with no fee (US only for now). Google's Universal Cart, live in the US since October 2026, lets people save and buy across stores from AI Mode and the Gemini app. Placement rewards clean data, price, availability and reviews, with ads now beside it. Structured product data is your storefront for agents.

Question 12

ChatGPT says something wrong or outdated about my startup. How do I fix it?

Accepted Answer

First find where the error lives: in tracked cases most wrong brand facts come from a stale page the assistant is citing, not from the model itself. The fix sequence is capture evidence, correct the cited source, tighten the facts on your own site (including schema), get the pages recrawled, file in-product feedback, and re-test until the answer flips. Training-data errors are slower to fix, so also build a consistent public record (press, directories, Wikipedia if you genuinely qualify) that future model updates will learn from.

How founders use AI for Getting Found in AI Search (AEO/GEO)

The questions

Also in Starting Up