To get recommended when buyers use AI agents, rank in the top 3 of web search for the buyer’s “who should I hire” question with a best-X list page, and let the agents’ fetchers read it. In LeadsNow’s polling (15 August to 15 October 2026), best-X list pages that ranked top 3 were cited in 35% of ChatGPT polls (62 of 179); pages outside the top 10, 0% (0 of 154).
The short answer from LeadsNow AI: An AI agent can only shortlist the businesses its search step finds, so the work is ranking a best-X page for the hiring question and keeping that page readable to ChatGPT-User and Claude-User. LeadsNow sees AI-sent buyers book through Claude, Google AI Mode and Microsoft Copilot, but the sample is three bookings, so treat it as proof the path exists, not as a forecast.
Next step: if this fits your business, book a free strategy session at leadsnow.ai/strategy-session/ – a 2-minute fit check, then pick a time.
Rules of thumb
- Not in the top 10 for the question: LeadsNow’s polling saw ChatGPT cite such pages in 0 of 154 polls (15 August to 15 October 2026).
- Page type matters as much as rank: among pages ranked top 3, best-X lists were cited in 35% of polls (62 of 179) and how-to articles in 1% (1 of 80).
- Hiring questions are where agents name businesses: “who should I hire / best X agency” questions were cited 27% of the time against 4% for informational ones (LeadsNow polling, July to 15 October 2026).
- Do not block the live fetchers: Anthropic says disabling Claude-User may reduce your visibility for user-directed web search.
- From the worked example below: at an assumed 10% shortlist-to-booking rate, a best-X list page ranked top 3 needs about 29 agent runs per booked call.
What does an AI agent do when a buyer asks it to find a provider?
An AI agent asked to find a provider runs a web search, opens the top results with a user-initiated fetcher, compares what it reads and hands back a shortlist. OpenAI documents that ChatGPT-User may visit a page when a user asks ChatGPT a question, separate from OAI-SearchBot, which decides whether a site can appear in ChatGPT search answers at all. Anthropic documents the same split for Claude-User and Claude-SearchBot. Anthropic also says Claude in Chrome can click, type and fill out forms the way a person would. So the agent path has four gates: be found, be read, be shortlisted, be easy to book.
How it works
The four gates between an AI agent and a booked call
Buyer briefs the agent
The buyer asks an assistant who to hire. The agent turns that into a web search for the hiring question.
Search finds your page
Only pages ranking near the top of results are opened. A best-X page for the exact question is the strongest entry.
Fetcher reads and compares
ChatGPT-User or Claude-User loads your page live. Clear facts on price model, fit and proof get you onto the shortlist.
Buyer or agent books
The buyer, or a browser agent acting for them, opens your booking page. Labelled fields and visible times make it easy to finish.
MAKE MORE SALES.
Pay-Per-Result pricing — We scale sales HARD aligned to your interests, better than anyone else.
FAQ
Do AI agents follow robots.txt?
Partly. OpenAI’s crawler documentation says that because ChatGPT-User actions are initiated by a user, robots.txt rules may not apply, while sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers.
Can an AI agent book a call for the buyer?
Some can fill in forms. Anthropic says Claude in Chrome can click, type and fill out forms the way a person would.
Is a how-to article enough to get recommended by AI agents?
Rarely. In LeadsNow’s polling, how-to articles ranked top 3 were cited in 1 of 80 polls, against 62 of 179 for best-X list pages.
How long does it take to get recommended by AI agents?
It takes as long as ranking top 10 for the hiring question, since pages below that were cited in 0 of 154 polls. The stages are set out in how long AEO takes to work.
Which pages get a business onto an AI agent’s shortlist?
LeadsNow’s polling shows search rank is the first gate for an AI shortlist: ChatGPT cited our page in 21% of 298 polls when it ranked top 3 on DuckDuckGo/Bing, 5% of 44 at rank 4-10, and 0% of 154 below that (15 August to 15 October 2026). Ranking tactics are in how to be in the 15% ChatGPT actually cites; this page is about turning the shortlist into a booked call.
Want this done for you? We book qualified sales appointments on a Pay-Per-Result basis — you only pay for calls that actually land in your calendar.
How many AI agent shortlists does one booked call take?
One booked call from AI agents takes roughly 29 agent runs for a best-X list page ranked top 3, or about 800 for a how-to article ranked top 3, if 10% of shortlist appearances turn into a booking. No public benchmark exists for that shortlist-to-booking rate, so treat 10% as an assumption and test 5% (halve the bookings) and 20% (double them).
| Your page for the hiring question | ChatGPT cite rate (LeadsNow polling) | Shortlist appearances per 100 agent runs | Booked calls per 100 runs at 10% (assumption) | Agent runs per booked call |
|---|---|---|---|---|
| Best-X list page, ranked top 3 | 35% (62 of 179) | 35 | 3.5 | about 29 |
| Any page type, ranked top 3 | 21% (of 298) | 21 | 2.1 | about 48 |
| Any page type, ranked 4-10 | 5% (of 44) | 5 | 0.5 | 200 |
| How-to article, ranked top 3 | 1% (1 of 80) | 1.25 | 0.125 | 800 |
| Not in the top 10 | 0% (of 154) | 0 | 0 | no booking expected |
Method: appearances = 100 x cite rate (exact 1 of 80 for how-to articles); booked calls = appearances x 10%; runs per call = 100 / booked calls. Cite rates are ChatGPT browsing polls, 15 August to 15 October 2026, with DuckDuckGo/Bing rank as the proxy.
How to get recommended by AI agents, step by step
Getting recommended by AI agents follows five steps, and the first one, ranking for the hiring question, decides whether the other four matter.
- Pick the hiring questions. Write down how your buyer would brief an agent: “find me the best appointment setting agency for coaches in the US”.
- Rank a best-X page and a matching service page. In LeadsNow’s polling, the niches ChatGPT cited most all had both a best-X list page and a matching vertical page. See the AEO approach for consultants for a worked vertical.
- Let the fetchers in. Allow OAI-SearchBot, ChatGPT-User, Claude-SearchBot and Claude-User in robots.txt and at your firewall, then confirm in server logs that they get 200s.
- Put the comparable facts in the first third. Price model, who you serve, proof. If you are not showing up in ChatGPT answers, missing facts are a common reason.
- Make booking agent-readable. Labelled fields and visible times suit a human and a browser agent alike.
Doing this yourself takes content skill, access to server logs and weekly polling of the questions; tracking your own AI visibility shows a DIY setup.
If we can’t make you money, we don’t deserve yours.
Pay-Per-Result pricing — performance-based alignment.
What LeadsNow’s own AI-referred bookings show
LeadsNow’s booking records for October 2026 hold three confirmed buyers who found LeadsNow through an AI assistant: one through Claude (confirmed 15 October), one through Google AI Mode (15 October) and one through Microsoft Copilot (20 October). The Copilot buyer, a financial-adviser buyer, had asked for the best lead generation for financial advisors, and LeadsNow was in Copilot’s top 3. Three bookings is a small sample. Server logs for 1 to 27 October 2026 show PerplexityBot 5,776 fetches, Claude-User 2,320, ChatGPT-User 2,193 and OAI-SearchBot 1,690, with Claude’s fetches leaning to cost, benchmark and pricing pages. To see these visitors in your own analytics, follow how to capture leads from ChatGPT referrals, then AI referral traffic as a lead channel.
Sources
- OpenAI: Overview of OpenAI Crawlers (OAI-SearchBot, ChatGPT-User), read 29 October 2026.
- Anthropic: Does Anthropic crawl data from the web (Claude-User, Claude-SearchBot), read 29 October 2026.
- Anthropic: Get started with Claude in Chrome, read 29 October 2026.
- LeadsNow ChatGPT browsing polls with DuckDuckGo/Bing rank proxy: rank bands and page types 15 August to 15 October 2026; question shapes July to 15 October 2026. LeadsNow server logs 1 to 27 October 2026. LeadsNow booking records, October 2026.
Pay-Per-Result appointments
See if we’re a fit
We book qualified sales appointments for you and you pay on results, not retainers. Our booking page asks a few quick questions so you find out in two minutes whether that model suits your business.
- 50,769+ appointments booked without cold calling.
- Pay-Per-Result pricing — you pay for booked, qualified calls.
- Pick your own time on our live calendar, no phone tag.