Mortgage & Lending AI Calling Compliance Guide (TCPA & State Rules)

Mortgage and lending AI calling is legal in 2026 when you obtain prior express written consent for autodialed or prerecorded/artificial-voice calls, honor federal and state Do-Not-Call rules, and disclose that the caller is an AI where required. The FCC has confirmed that AI-generated voices fall under the same TCPA restrictions as artificial and prerecorded voices, so a consent gap that was risky before is now unambiguous liability. This matters because the average B2B lead sits for roughly 29–47 hours before a first response, and studies show around 78% of buyers purchase from whoever contacts them first — meaning compliant speed is directly tied to your funded-loan revenue.

This playbook walks lenders, brokers, and mortgage marketers through the consent standard, state overlays, AI-voice disclosure, and an operational checklist for calling inbound leads in seconds without tripping TCPA.

The short answer: what makes an AI mortgage call compliant

A compliant AI mortgage call requires documented prior express written consent, DNC compliance, and honest disclosure of automation — plus a clean audit trail.

The Telephone Consumer Protection Act (TCPA) governs how you can dial consumers with automated technology. For mortgage and lending calls that use an autodialer, a prerecorded message, or an artificial/AI voice, the standard is prior express written consent (PEWC) because the calls are telemarketing to a personal number.

The core requirements break down to:

• Consent before the call, not captured during it. The consumer must agree in writing to receive autodialed/AI calls at a specific number.
• Clear and conspicuous disclosure at the point of consent that they'll receive automated or prerecorded calls, and that consent isn't a condition of purchase.
• Do-Not-Call scrubbing against the National DNC Registry and your internal list.
• Identification of who is calling and on whose behalf.
• Time-of-day limits (generally 8 a.m.–9 p.m. in the consumer's local time under federal rules; many states are stricter).

The FCC's 2024 declaratory ruling made explicit that AI-generated voices are "artificial" voices under the TCPA — so an AI calling agent needs the same PEWC as a robocall. Treat your AI dialer as a prerecorded-voice system for consent purposes, and you stay inside the rules.

Prior express written consent is the whole ballgame

Prior express written consent is the single most important compliance control for AI mortgage calling — get it right and most TCPA exposure disappears.

PEWC is a specific legal term, not just a checkbox. To qualify, the consent language must:

• Be a written agreement (a web form, e-signature, or recorded verbal agreement can count) that the consumer signs or affirmatively submits.
• Clearly authorize the seller to deliver advertisements or telemarketing messages using an autodialer or artificial/prerecorded voice to the number provided.
• State that consent is not a condition of purchasing any goods or services.
• Identify the specific seller who will be calling — a shared "our partners" clause is weak and increasingly litigated.

For lead-gen mortgage funnels, this is where deals break. If you buy leads from an aggregator, the consent was collected by someone else. Under the newer one-to-one consent expectations regulators have signaled, a single form consenting to dozens of unnamed "marketing partners" is fragile. Verify that the consent names your company, or is transferable, and that you can produce the exact disclosure text, timestamp, IP, and the number consented to.

Store all of this. In litigation, the burden is on you to prove consent — not on the consumer to prove they didn't give it. A calling platform with a built-in CRM that logs recordings, transcripts, and consent metadata turns that burden into a quick export instead of a scramble.

The FCC AI-voice ruling changes your disclosure obligations

AI-generated voices are treated as artificial voices under the TCPA, which means AI mortgage calls need consent and, in many contexts, disclosure that the voice is not human.

The 2024 FCC declaratory ruling closed the debate: if your calling agent uses synthesized or cloned voice technology, it is subject to the same rules as a robocall. There is no "the AI sounds human, so it's like a live agent" loophole.

Practical implications for lenders:

• You need PEWC for AI-voice outreach to a consumer's cell or residential line, just as you would for a prerecorded message.
• Disclose the automated nature early in the call. Several states and pending rules push toward requiring callers to disclose that the consumer is speaking with an AI. Leading with "Hi, I'm an automated assistant from [Lender], calling about the rate quote you requested" is the safe posture.
• Offer a fast path to a human. A compliant AI agent should qualify and then warm-transfer to a licensed loan officer, not trap the borrower in a bot loop.

The safest architecture is AI-for-speed, human-for-substance: the AI reaches the inbound lead in seconds, confirms interest, and hands a warm, licensed human the conversation about actual loan terms. That also keeps you clear of licensing issues, since mortgage advice and rate quoting generally require a licensed originator (NMLS).

State rules stack on top of federal TCPA — and some are stricter

State telemarketing and "mini-TCPA" laws add consent, registration, and calling-window requirements beyond the federal baseline, and several carry their own private lawsuits.

Federal TCPA is the floor, not the ceiling. A partial map of the stricter overlays lenders hit most often:

State	What's different	Watch-out for lenders
Florida (FTSA)	Mini-TCPA covering autodialed/automated calls & texts; private right of action	Aggressive plaintiff activity; tighter consent expectations
Oklahoma	Broad mini-TCPA with statutory damages per violation	Applies to automated sales calls to residents
Washington	Commercial call disclosure & do-not-call rules	Identification and opt-out handling
Texas	Registration requirement for telephone solicitors	May require a filing/bond before dialing
Multiple states	Narrower calling windows (e.g., start after 8/9 a.m., stricter Sunday/holiday limits)	Always use the consumer's local time

This table is a general summary; state statutes and enforcement change frequently — verify current requirements with counsel before launching a campaign.

Three operational rules keep you clean across states:

• Call in the consumer's local time zone, and default to the stricter of federal vs. state windows.
• Scrub against state DNC lists where they exist, in addition to the federal registry.
• Check registration/bonding requirements in states like Texas before you dial residents there.

Because mini-TCPA statutes often carry a private right of action with per-call statutory damages, a single misconfigured campaign can generate class exposure. Geo-aware calling logic isn't a nice-to-have — it's the difference between a scalable program and a lawsuit magnet.

Speed and compliance are not in tension — they reinforce each other

The fastest compliant path to a lead is to call the instant consent is captured, because that's exactly when TCPA risk is lowest and conversion is highest.

The most defensible call you can make is to a fresh inbound lead who just submitted a form consenting to be contacted. The consent is timestamped, the interest is unambiguous, and the consumer expects to hear from you.

The revenue case is overwhelming:

• The MIT/Oldroyd Lead Response Management study found leads contacted within five minutes are dramatically more likely to qualify — the widely cited figure is around 21x versus waiting 30 minutes.
• Velocify research shows contact within the first minute drives materially higher conversion.
• Roughly 78% of buyers purchase from the first responder, and 30–40% of inbound leads arrive after business hours when human teams are offline.

That last stat is why AI calling exists: a human team can't answer a 11:40 p.m. rate-quote form in 10 seconds, but a compliant AI agent can — then warm-transfer or schedule for the morning. Contacting a consumer minutes after they consented is far safer than a batch dial to a two-week-old purchased list. For the full framework on why response velocity drives funded loans, see the complete guide to speed to lead, and the mechanics of how AI calling works.

Your mortgage AI calling compliance checklist

Use this operational checklist to launch AI calling that survives a TCPA audit.

Before you dial:

• Capture PEWC that names your company, discloses automated/AI calls, and states consent isn't a purchase condition.
• Log the consent text, timestamp, IP/source, and the exact number.
• Scrub every number against the federal DNC Registry, applicable state lists, and your internal do-not-call/opt-out list.
• Confirm any state registration/bonding (e.g., Texas) before calling those residents.

During the call:

• Disclose the caller's identity and the lender on whose behalf you're calling.
• Disclose the automated/AI nature of the voice.
• Honor calling windows in the consumer's local time (default 8 a.m.–9 p.m., stricter where state law requires).
• Warm-transfer rate and product questions to a licensed (NMLS) loan officer.
• Process opt-out requests immediately and permanently.

After the call:

• Retain recordings, transcripts, and AI summaries for your retention window.
• Suppress opted-out numbers across all campaigns within the required timeframe.
• Re-verify consent age and DNC status before any re-contact.

Vendor due diligence — when choosing an AI calling platform, confirm it can: enforce time-zone-aware calling windows, ingest and honor DNC/opt-out lists, disclose AI status, warm-transfer to humans, and export a complete consent-and-call audit trail. If a vendor can't produce that record on demand, the compliance risk is yours, not theirs.