
What you can do today
- Put your church name, address and service times on your own website as plain text. It is the one page you control.
- Find your church in your denomination's own finder and in Church Finder. Fix a wrong address or time. People use these, even if no one has measured ChatGPT doing so.
- Keep Google, Yelp, Facebook and Apple Maps identical on name, address and service times.
- Don't pay for a directory package because someone says it helps AI. I found no measurement that it does.
- Check that your own site lets the robot in, and look again in six months.
What to put on that page: a Plan Your Visit page.
What the files say

I found no measurement of which directories ChatGPT uses for churches. What I can show is which let in OAI-SearchBot, OpenAI's search robot, the program that reads pages for ChatGPT answers.
A site's robots.txt, a small file that asks robots which pages to skip, says so. I read each one on November 16, 2026. Files change.
- Let in: Church Finder, UCC (ucc.org), UMC, SBC, 9Marks, church.org, FaithStreet, The Gospel Coalition
- Mostly let in: TripAdvisor, BBB
- Articles only: Yelp
- Kept out: MapQuest, Facebook, Nextdoor, Apple Maps website
The details:
- Church Finder: only admin, search, login, claim and review pages closed.
- UCC: blocks GPTBot, OpenAI's training robot, but not the search robot.
- UMC, SBC, 9Marks, church.org, FaithStreet, The Gospel Coalition: no rule about AI robots.
- TripAdvisor: search, account, user-review and some country pages closed.
- BBB: review and complaint pages, and web addresses with a ?, closed.
- Yelp: business pages closed.
- Nextdoor: only two small paths open.
findachurch.com and yellowpages.com returned errors (503 and 403) to my request, so I could not read their files.
A request, not a lock
A robots.txt is a request. Honest robots follow it, and it stops no one else. OpenAI says ChatGPT-User, the robot that fetches a page when a person asks about it, may ignore it: "Because these actions are initiated by a user, robots.txt rules may not apply."
A company can also hand its listings to OpenAI under a licence or a data feed, a direct supply of its data. That never touches robots.txt. I did not find out which of these directories do.
What people do with directories
In Lifeway Research's 2023 survey, 15% of people who changed churches first found theirs "on the internet." Whether directories matter to ChatGPT is a separate question, and I found no study of it.
Based on public documentation: OpenAI documentation; Lifeway Research; robots.txt of churchfinder.com; robots.txt of ucc.org; robots.txt of umc.org; robots.txt of sbc.net.
- OpenAI documentation: crawlers (OAI-SearchBot, ChatGPT-User)
- Lifeway Research, 2023-11-09: Church switchers identify factors behind their decision
- robots.txt of churchfinder.com, read 2026-11-16
- robots.txt of ucc.org, read 2026-11-16
- robots.txt of umc.org, read 2026-11-16
- robots.txt of sbc.net, read 2026-11-16
- robots.txt of 9marks.org, read 2026-11-16
- robots.txt of church.org, read 2026-11-16
- robots.txt of faithstreet.com, read 2026-11-16
- robots.txt of thegospelcoalition.org, read 2026-11-16
- robots.txt of tripadvisor.com, read 2026-11-16
- robots.txt of bbb.org, read 2026-11-16
- robots.txt of yelp.com, read 2026-11-16
- robots.txt of mapquest.com, read 2026-11-16
- robots.txt of facebook.com, read 2026-11-16
- robots.txt of nextdoor.com, read 2026-11-16
- robots.txt of maps.apple.com, read 2026-11-16
