Here is something I hear from plaintiff firms more often than I should: "Our admin handles outreach. We don't worry about the compliance piece."
That sentence used to be mostly fine. When outreach meant a mailed notice and a follow-up postcard, the regulatory surface area was small. But the moment your settlement administrator picks up the phone, sends an SMS, or deploys an AI voice agent to reach your 8,000 claimants, you are in TCPA territory. And the FCC has made one thing very clear in the last twelve months: AI-generated voices are "artificial or prerecorded" under the statute.
That classification matters. A lot. Because the liability for a non-compliant call does not stop at the vendor who placed it. It follows the chain back to the entity that authorized it. Which, in a mass arbitration matter, is your firm.
What the FCC Actually Said
The FCC's declaratory ruling classified AI-generated voice as falling squarely under the TCPA's restrictions on "artificial or prerecorded voice" calls. That means any outbound call using AI voice technology to reach a consumer requires prior express consent for informational calls, or prior express written consent for anything that could be construed as marketing or solicitation.
The penalties have not changed: $500 per violation, up to $1,500 if willful, with no aggregate cap. On a 10,000-claimant matter, even a narrow compliance failure across a single outreach wave could generate seven-figure exposure before anyone files a motion.
And the revocation rules tightened alongside this. Consumers can now revoke consent through any reasonable method (a text reply, an email, a voicemail, a conversation with a live agent), and your system has to honor that revocation within 10 business days. Not your vendor's system. Your system. Because the consent chain traces back to you.
Where the Gap Opens
Most legacy claims administrators were not built to manage TCPA-grade consent at the claimant level. They were built to send court-approved notices. The compliance framework they inherited is Rule 23 notice adequacy, not FCC consent management.
When those same administrators bolt on AI voice or automated SMS to improve completion rates (which is the right instinct, by the way), they often do it through third-party vendors. A dialer platform here, an AI voice API there, a separate SMS gateway. Each vendor has its own consent assumptions. And none of them are issuing your firm a clean audit trail that maps consent-to-contact back to a specific claimant, a specific channel, and a specific timestamp.
The problem is not that AI outreach is bad. It is genuinely one of the best tools available for lifting completion rates across a 180-day window. The problem is that most firms have no visibility into how that outreach is being conducted on their behalf.
Ask your current administrator three questions:
- For every AI voice call or automated text placed to a claimant, can you produce the consent record (timestamp, source, exact disclosure language) within 48 hours?
- When a claimant replies "stop" to a text or tells an AI agent they want to opt out, how many business days until that revocation is honored across all channels?
- Is your AI voice system classified internally as "prerecorded" for TCPA purposes, and does your consent capture reflect that classification?
If the answer to any of those is vague, you have a gap. And the gap is yours, not theirs.
Why Owning the Stack Changes the Compliance Picture
This is the part where I'm biased, so I'll say it upfront: I built a stack, and I think owning the technology changes the math. Here's why.
When your document signing, SMS delivery, email delivery, AI voice agent, and consent management all live on the same platform, you get one thing that a five-vendor assembly cannot give you: a single, unified consent chain per claimant.
At GroupSettle, when a claimant signs their fee agreement through Send It By Text's native document signing, that signature event creates the consent record. When outreach goes out via SMS or email, it references that record. When the AI super agent (which handles over 80% of inbound claimant inquiries on its own) takes a call, it checks consent status in real time. When a claimant revokes, it propagates across every channel instantly, not in 10 business days, not after a sync cycle between three vendors.
That is not a feature. It is an architecture decision. And it is the kind of decision that matters when a defense firm files a motion challenging the procedural integrity of your claimant outreach, which is a motion you should expect to see more of in 2026 and 2027.
The Defense Bar Is Watching This Closely
If you have read the defense-side commentary coming out of firms like Troutman and Duane Morris, you already know: challenging the legitimacy of claimant communications is becoming a standard play in mass arbitration. Not just "are these real claimants?" but "were these claimants contacted lawfully?"
A process arbitrator who sees that 6,000 claimants were reached via AI voice calls, and that the plaintiff firm cannot produce individualized consent records for each, has a real procedural question on their hands. That question can delay proceedings, reduce your filed count, or give the defense ammunition for a sanctions motion.
The firms that are winning this part of the game are not the ones avoiding AI outreach. They are the ones who can show the arbitrator a clean record for every single touchpoint. Consent captured at signing. Channel-specific authorization. Real-time revocation honored. Full audit trail.
The Practical Audit You Can Run This Week
You do not need to switch vendors to start protecting yourself. Here is what you can do before Friday:
- Pull your settlement administration agreement and read the indemnification clause. Who bears liability for TCPA violations arising from claimant outreach? If it is silent, that silence probably does not favor you.
- Request a sample consent chain from your administrator for 10 random claimants. You want: the consent capture event, the disclosure language shown, the timestamp, and the channel authorization. If they cannot produce this within a week, that tells you something.
- Ask whether AI voice or automated dialing technology is used at any point in the outreach sequence. If yes, confirm that the consent capture specifically covers "artificial or prerecorded voice" calls, not just generic contact authorization.
- Check revocation handling. Send a test opt-out through each channel (SMS reply, email, voicemail) and measure how long it takes to propagate.
These are not theoretical exercises. They are the questions a process arbitrator or a defense firm will eventually ask. Better to have the answers now than to discover the gaps under a motion.
Completion Still Requires Persistence. Persistence Requires Compliance.
The whole thesis behind mass arbitration completion is that you need to keep showing up. The claimant who ignores week one will sign in week twelve if you are persistent and present across channels. That thesis is correct. But persistence without a defensible consent infrastructure is just exposure.
The firms that will hit their 85% or 90% release thresholds consistently are the ones that built their outreach on a compliance-first foundation. Not because they are cautious by nature, but because they understand that every touchpoint is either an asset (a documented, consent-backed engagement) or a liability (an unaudited call that a defense firm can weaponize).
Build the asset. Document the chain. Own the record. That is how persistence becomes your competitive advantage instead of your biggest risk.
This is the kind of compliance and completion analysis GroupSettle runs for plaintiff firms before a single outreach message goes out. If you want to see how a unified stack handles consent at the claimant level, reach out to Kasia at (813) 737-7025 or visit massarb.groupsettle.com.