10DLC Registration for Insurance Agents: Setup Guide

Here is the failure mode that costs agents the most and gets noticed the least. You text a lead a confirmation for tomorrow's appointment. Your system shows the message as sent. There is no error, no bounce, no red flag anywhere in your dashboard. The message simply never arrived on the prospect's phone, and you find out the next day when nobody picks up.

That is carrier filtering, and the reason it happened is almost always the same: the number you sent from was not registered under A2P 10DLC, or it was registered under a campaign that does not match what you actually sent.

This is the least glamorous piece of an agent's tech stack and one of the few where a bad setup is invisible. A blocked call is obvious. A filtered text looks exactly like a delivered one from your side of the glass.

What 10DLC actually is

A2P 10DLC stands for application-to-person messaging over ten-digit long codes — ordinary local phone numbers, the kind you would recognize as a normal number. “Application-to-person” means the message originates from software rather than a human thumbing a keyboard on a handset.

For years, businesses sent bulk text from regular local numbers and carriers had no way to tell that traffic apart from person-to-person messaging. Spam volume made that untenable. So the US carriers built a registration system: declare who you are, declare what you intend to send, attach your numbers, and get scored.

Two things follow from that score. First, throughput — how many messages per second you are allowed to push. Second, and far more important for an agent, deliverability — how hard the filters lean on your traffic before it reaches a handset.

Notice what this is not. It is not a law and it is not a license. No regulator issues it and no state board checks it. It is a private gate run by the companies that own the pipes, which means the practical consequence of ignoring it is not a fine, it is silence.

Why insurance gets extra scrutiny

Every use case gets reviewed. Some get reviewed harder, and insurance sits in that group alongside lending, debt relief, and anything else that touches a consumer's money.

The reason is straightforward. These verticals generate the most consumer complaints, the most spam reports, and the most litigation. Reviewers have seen thousands of filings from agents texting purchased lists, and your application is read against that backdrop whether or not it is fair to you.

What that means practically: assume your opt-in evidence gets read closely, assume your sample messages get read closely, and do not submit anything vague hoping it slides through. It will come back, and each round trip costs days.

The three layers you register

Registration is not one filing. It is three nested objects, and agents get confused because platforms name them differently. The structure underneath is always the same.

The third layer is where a surprising number of agents fall down after doing the first two correctly. You get approved, then you buy a new number a month later, forget to attach it, and that number sends unregistered traffic from day one. If your dialer rotates numbers for caller ID, every number in the rotation needs to be attached — which is one more argument for a platform that manages the whole pool rather than leaving you to track it in a spreadsheet.

What to have ready before you start

Gather this first. Submitting halfway and guessing at the rest is how filings get rejected on details that had nothing to do with your messaging.

• Your legal business name and EIN, exactly as the IRS has them.Not your DBA, not the name on your business cards. Verification is automated against tax records and it is unforgiving about punctuation, abbreviations, and “LLC” versus “L.L.C.”
• A registered business address. The one on the EIN filing.
• A working website that describes what you do. A one-page site is fine. A parked domain, a Facebook page, or nothing at all is a common rejection.
• The exact opt-in flow, documented. A live URL to the form, or a screenshot showing the checkbox and the disclosure text next to it. This is the single most scrutinized item.
• Two or three real sample messages. Written the way you would actually send them.
• Your privacy policy and terms. Published and reachable, with your messaging practices mentioned.

If you are a sole proprietor without an EIN, most platforms offer a reduced sole-proprietor path with much lower throughput and tighter limits. It works for light appointment confirmations. If texting is going to be a real part of how you work leads, getting an EIN is the cleaner road.

The opt-in disclosure is where agents fail

If your campaign gets rejected once, this is probably why. Reviewers are not asking whether you have consent in some general sense. They want to see the exact moment a consumer agreed, and they want the words that were on the screen when it happened.

A disclosure that passes review generally contains all of these:

• Your business name. The one on the Brand registration, not a marketing alias.
• What kind of messages they will get. Quotes, appointment reminders, policy follow-up.
• That message and data rates may apply.
• How to stop. Reply STOP to opt out, reply HELP for help.
• A link to your privacy policy and terms.
• A statement that consent is not a condition of purchase. This one gets skipped constantly and it matters on both the carrier side and the legal side.

One structural rule worth internalizing: the consent to receive texts has to be its own affirmative act. Not bundled into a general terms checkbox, not pre-ticked, not buried in a paragraph nobody reads. A separate, unchecked box with the disclosure sitting right next to it.

That standard is not arbitrary — it is close to what a defensible TCPA consent record requires anyway. Building it once satisfies both. Which is the useful way to think about this whole exercise: the carriers are asking to see the same evidence a plaintiff's attorney would ask for.

Sample messages: what passes, what comes back

Sample messages are the other half of the review. They get read as evidence of intent, so write them the way you will actually send them.

A version that clears review looks closer to this:

Business name, a specific reference to why you are contacting them, and an opt-out. Include the opt-out on at least one sample even if you do not append it to every real message, because the reviewer is checking that you know the requirement exists.

Use your own domain for links rather than a public shortener. Shared shorteners are among the most filtered patterns in messaging, for the obvious reason that spammers rely on them.

What it costs

Fees are set by the registry and the carriers, so your platform passes them through rather than setting them. The structure is stable even when the exact numbers move.

For a solo agent sending appointment confirmations and follow-ups, the fixed cost is not the deciding factor. Confirm current amounts with your provider before you budget, since the registry adjusts them.

10DLC approval is not TCPA consent

This deserves its own section because the confusion is common and expensive.

An approved campaign means carriers have agreed to carry your traffic. It says nothing about whether you were legally permitted to message any particular person. Those are different questions answered by different systems.

Read the last row twice. A rejected campaign is an inconvenience. A text sent without prior express written consent is exposure you cannot retroactively fix, at statutory damages per message. The TCPA rules for texting leads are the layer that actually decides whether you get sued.

The rejections that come back most

• Business details do not match tax records. Almost always a name formatting difference. Pull the actual EIN letter and copy it character for character.
• Opt-in evidence is a description instead of proof.Writing “consumers opt in on our website” is not evidence. A URL to the live form, or a screenshot with the disclosure visible, is.
• The website does not support the stated use case. If the campaign says appointment reminders and the site says nothing about insurance services, that mismatch gets flagged.
• Sample messages omit the business name or opt-out. Cheap to fix, and the most frequent single cause.
• Use case reads as cold outreach. Describing purchased or aged lists in your campaign narrative invites rejection. If you are working a list you did not generate, sort out the consent question before the filing question.

Rejections are not permanent. You fix the item and refile. But each cycle runs days and the fees are generally not refunded, so it is worth getting right on the first submission.

After approval: staying in good standing

Approval is a starting position, not a permanent state. Carriers keep scoring your traffic afterward, and the things that degrade a score are behavioral.

• Honor STOP instantly and permanently. Platforms handle this automatically. If yours does not, that is a serious problem well beyond 10DLC.
• Do not send outside your registered use case. Approved for appointment reminders and then blasting promotions is exactly what suspension is for.
• Keep new numbers attached. Every number added to your pool needs a campaign assignment.
• Watch complaint and opt-out rates. A rising opt-out rate is telling you the list or the message is wrong, before the carriers tell you louder.
• Do not spread identical traffic across many numbers to dodge limits. Snowshoeing is well understood by carrier analytics and it is the fastest way to poison a whole pool.

That last habit is the messaging cousin of the caller-ID problem behind spam-likely flags on your outbound calls. Both are reputation systems, both score behavior rather than intent, and both punish the same pattern: high volume, low engagement, spread thin across numbers.

What if you only call and never text

Then 10DLC does not apply to you. It is a messaging framework, full stop.

Voice has its own parallel systems — STIR/SHAKEN attestation, carrier analytics, and the number reputation scoring that produces spam labels. Different mechanism, same underlying logic: carriers decide how much of your traffic to trust.

Plenty of agents run voice-only and do fine. The trap is the middle position — deciding to add texting later, sending a few from an unregistered dialer number to test it, and concluding texting does not work for your market when what actually happened is that nothing was delivered. If you are evaluating dialers, ask directly whether the platform registers 10DLC on your behalf, whether it handles brand vetting, and whether new numbers are auto-attached to your campaign. A vendor that hands you a registry login and wishes you luck is handing you a project.