California Attorney Advertising Rules and Automated Text Intake

Last reviewed 31 July 2026  ·  Written for personal injury firms switching on automated text intake in California.

Does replying to an inbound lead count as solicitation in California?

No. California Rule of Professional Conduct 7.3¹² defines solicitation as a targeted communication initiated by or on behalf of the lawyer. Comment [1] to the rule states that a communication is not a solicitation if it is made in response to a request for information. Someone who submits your contact form has made that request.

Rule 7.3(e) sets out the definition:

The word doing the work is initiated. An automated reply to a form submission or a missed call is not initiated by the lawyer in the sense the rule means. It is the second message in a conversation the other person started. That is why Comment [1] carves out communications made "in response to a request for information."

Get this backwards and the analysis changes completely. Buying a list of accident reports and texting the people on it is a targeted communication that you initiated, to a specific person, offering legal services. That is squarely inside the definition, and paragraph (a) is where it starts running into trouble.

Which parts of Rule 7.3 actually apply to a text message?

Paragraph (a) bans in-person, live telephone and real-time electronic solicitation for pecuniary gain. Paragraph (b) bans any solicitation of someone who has said they do not want to be solicited, and any solicitation involving intrusion, coercion, duress or harassment. Paragraph (c) requires the word "Advertisement" on solicitation communications².

Paragraph (a) is about real-time contact. A text message is generally not treated as real-time electronic contact in the way a live call or a video chat is, since the recipient can ignore it. But paragraph (b) applies to written and electronic communication regardless, which is why a working opt-out is not a nice-to-have.

Paragraph (c) is the labeling rule:

Note the trigger: it applies to a communication soliciting professional employment. If the communication is not a solicitation because the person asked you to contact them, paragraph (c) is not engaged. That is the whole practical difference between an inbound intake reply and an outbound campaign.

Side by side, the two situations diverge almost completely:

The bottom four rows are the point. Nothing about being inbound switches off Rule 7.1 or the opt-out obligation. What being inbound switches off is the machinery built to stop lawyers approaching strangers.

Does California require you to file your advertising with the State Bar?

No. Rule 7.2³ permits advertising through any written, recorded or electronic means, subject to Rules 7.1 and 7.3. It contains no filing requirement, no pre-approval process, and no record-retention obligation. California removed the older two-year retention rule when the current rules were adopted.

This is a genuine difference from Texas and Florida, both of which run filing programmes with fees and deadlines. California firms sometimes assume the same machinery exists here and either over-comply or panic about it. It does not.

That absence cuts both ways. Nobody is going to pre-clear your intake script for you, so the responsibility for it sitting inside Rule 7.1 sits entirely with the firm. Rule 7.1¹ is short and absolute: a lawyer shall not make a false or misleading communication about the lawyer or the lawyer's services. An automated message that implies you have already evaluated someone's case, or that suggests an outcome, is a Rule 7.1 problem no matter how fast it arrived.

What should an automated California intake message never say?

Nothing that predicts an outcome, values a claim, comments on a deadline, or advises the person about dealing with an insurer. Nothing that implies a lawyer has reviewed the matter when no lawyer has. Nothing that suggests an attorney-client relationship has formed. And nothing the firm has not read and approved in writing.

Automated intake is also where the unauthorized practice of law question shows up, because the software is talking to a member of the public about their legal situation without a lawyer present. The safe design is narrow: the system collects facts and confirms contact details, and any question that calls for judgment gets a fixed response saying an attorney will answer it.

It is worth writing that deflection line out explicitly and approving it, rather than trusting a model to improvise. "That is a question for one of our attorneys. I have made a note of it and someone will call you back during business hours" is a complete answer, and it never becomes advice.

What does a compliant California intake script actually look like?

A compliant script is narrow and written down. It identifies the firm, says plainly that the sender is not an attorney, collects facts without characterising them, uses one fixed line whenever judgment is requested, and offers a working opt-out. Below is a complete first exchange with the reasoning for each line.

Take this as a starting draft to mark up with your own ethics counsel, not as an approved script. The annotations tell you what each line is doing, so you can tell what breaks if you change it.

The five lines to check before you approve anything

1. The not-an-attorney line. First message, before any question. It is the cheapest protection in the whole script.
2. The deflection line. Fixed wording, approved by the firm, triggered by any request for judgment.
3. The opt-out line. First message, plain English, and it has to actually work.
4. The closing line. A callback time and nothing more.
5. The Spanish version. If you serve Spanish-speaking clients, the Spanish script needs the same review as the English one. A translated deflection line that softens into advice is the usual failure.

Read the whole thing out loud in the recipient's position before you sign it off. Anything that sounds like an opinion when spoken is an opinion when read.

How does the TCPA change the picture for California firms?

The federal Telephone Consumer Protection Act applies on top of the bar rules and does not care whose rules you followed. Statutory damages are $500 per message, or $1,500 per message for a willful or knowing violation⁵, so exposure scales with volume rather than with intent.

For inbound intake the posture is comparatively good. The person handed you their own mobile number and asked you to contact them, and since Facebook v. Duguid⁶ the definition of an automatic telephone dialing system has been read narrowly enough that a one-to-one reply triggered by a human's own form submission is usually outside § 227(b)'s core prohibition.

The part that is not optional is revocation. Since 11 April 2025, FCC rules require that a consumer be able to revoke consent using any reasonable wording, and that the revocation be honoured within ten business days. The wider "revoke all message types at once" element of that rule has now been pushed to 31 January 2027⁷ after a further extension, but the ten-business-day core is already live. Practically: STOP has to work, and so do the twenty other things people type instead of STOP.

A short checklist before you switch anything on

• Confirm the trigger is inbound only. The system should be physically incapable of texting a number that did not come from your own form, your own missed call, or your own tracked ad click.
• Read the whole script as the recipient. Every line, in order, out loud. If any line sounds like advice, rewrite it.
• Approve it in writing. Your name is on the message; the vendor's is not.
• Test STOP, plus "stop texting me", "remove me" and "quit". Then test that a follow-up genuinely does not send.
• Keep the transcripts. Rule 7.2 does not require it, but a complete log is the only thing that answers a complaint about what was said.
• Re-check the rule text before you rely on this page. The State Bar circulated proposed amendments to Rule 7.3 for public comment in 2025⁴.