How do SMS opt-out rules apply to AI texting?

Exactly as they apply to a person on your team. Automation does not change the obligation, and the volume automation makes possible is precisely why the obligation matters more.

The rules in outline

Messaging law generally turns on consent before you send, a working opt-out mechanism, and honouring that opt-out promptly. There are limits on timing and on automated dialling, and the specifics vary by state and by message type.

Marketing messages are treated more strictly than transactional ones, and the boundary between the two is less obvious than it sounds. A message about a newly available unit is closer to marketing than a maintenance appointment confirmation, even when both go to somebody you have an existing relationship with.

Why cross-channel syncing matters

This is the failure we see most often, and it rarely comes from bad intent. Somebody replies STOP to a text and is removed from the SMS list. Two weeks later the same person receives an email, or a call, from a different system that never heard about it.

From the customer's side that is one business ignoring them. Opt-out state is synced across SMS, WhatsApp, email and voice, so opting out anywhere means opting out everywhere.

What to get right at setup

• Record where consent came from and when, per contact, not as a blanket assumption.
• Treat inherited or purchased lists as suspect rather than consented.
• Make the opt-out instruction present and unambiguous, not buried.
• Keep sending inside permitted hours in the recipient's timezone, not yours.
• Log every send and every opt-out, because the record is what you would need if challenged.

The re-engagement trap

Messaging a dormant list is where rules get broken by accident, because the contacts are old and the consent behind them is often undocumented.

A contact who inquired two years ago and never heard from you since is not obviously a live consent. Before running any re-engagement, it is worth deciding how far back you are willing to reach and being able to say why, rather than messaging everything in the database because it is technically possible.

Not legal advice

This is a genuinely technical area and it changes. Your counsel should review your consent position and your message templates before you scale up volume. What the system gives you is consistent enforcement of whatever you decide, and a log that shows it.

Opt-out is not one word

A detail that catches people out in practice. Opt-out handling is often built around a keyword, and real people do not use keywords.

• Please stop texting me.
• Take me off this list.
• I already found somewhere, thanks.
• Wrong number.
• Do not contact me again.

Only one of those is likely to trip a keyword filter, and every one of them is a clear request to stop. A system that keeps messaging after the second or third of those has technically honoured the rule and has plainly ignored the person, which is not a distinction worth defending to anybody.

Wrong number deserves its own handling. It is not an opt-out from the contact you meant to reach, it is a signal that the number on that record is somebody else's, and continuing to use it is both useless and a nuisance to an uninvolved person.

Where consent actually came from

The question that decides most of this is one most operators have never had to answer precisely: for any given contact, what did they agree to and when.

Somebody who submitted a web form about one property, somebody who called you, somebody imported from a list a previous manager bought, and a current resident who gave a number on a lease are in four different positions. Recording the source and the date per contact is unglamorous work that takes the guesswork out of every later decision about who can be messaged.