AI marketing calls spur suit against lender: practical TCPA lessons for marketers
A Michigan mortgage lender is facing a class action alleging it used an AI voice agent to make outbound sales calls without proper consent. According to a complaint filed in the Eastern District of Michigan, Pennsylvania homeowner Brennan Landy received an unsolicited call in January that featured an artificial voice and lacked required disclosures.
The call, the complaint says, included awkward pauses, odd inflections, and stilted replies-signals of an automated or AI-driven agent. The agent eventually transferred Landy to a human employee, tying the outreach to Mortgage One Funding. The case seeks to represent others who received similar calls.
The case at a glance
- Who: Plaintiff Brennan Landy vs. Mortgage One Funding (Clawson, MI; licensed in 45 states, per NMLS).
- Where: U.S. District Court, Eastern District of Michigan.
- Allegations: Unsolicited AI-voice telemarketing to a cellphone, no upfront disclosure that an artificial voice was used, and calling a number on the National Do Not Call Registry-all violations alleged under the TCPA.
- Exposure: $500 per call; up to $1,500 per call if a court finds willful or knowing violations.
Why marketers should care
AI voice agents are tempting: lower costs, faster throughput, 24/7 availability. But one noncompliant campaign can turn into thousands of violations-plus headlines and distrust. There's also the risk of AI agents saying the wrong thing about rates, terms, or eligibility and handing you a compliance mess.
Even if you use a third-party vendor, you don't outsource liability. If it's your campaign, you own the risk.
What the TCPA expects from AI-driven calling
- Consent first: Marketing calls to cellphones with a prerecorded or artificial voice generally require prior express written consent.
- Disclose upfront: Identify the caller at the start, and if you're using an artificial or prerecorded voice, say so immediately.
- Honor Do Not Call: Scrub against national and internal DNC lists and provide a clear, automated opt-out.
- Keep receipts: Maintain timestamped consent records, disclosures, scripts, transfers, and opt-outs.
Reference materials: FCC: Telemarketing and robocalls and the National Do Not Call Registry.
Risk isn't just fines
- Carrier blocking and spam labeling if complaints rise.
- Reputational damage and lower answer rates across all channels.
- Misstatements from AI agents about products or pricing that create regulatory exposure and consumer harm.
Action checklist for marketing teams
- Consent capture: Use clear, unbundled language for AI/prerecorded outreach. Store proof (timestamp, source URL, IP, form snapshot).
- Opening script: "This call uses an AI-assisted voice on behalf of [Brand]. Press 2 to speak with a human now." Keep it short and obvious.
- DNC hygiene: Scrub against national and state lists. Maintain an internal DNC. Honor opt-outs within 24 hours.
- Quiet hours and frequency caps: Respect state call windows and limit attempts per lead. Track complaints per 1,000 calls.
- QA before scale: Test for awkward pauses, barge-in handling, and natural escalation to a human. Kill-switch ready.
- Content guardrails: No rate quotes or eligibility claims by AI. Route sensitive questions to licensed humans.
- Vendor diligence: Contract for TCPA compliance, data provenance, opt-out handling, disclosure requirements, and audit rights. Validate their scrubbing and consent flows.
- Incident response: Central log retention, immediate pause on affected campaigns, rapid review with counsel, and consumer remediation steps.
What happened, specifically
The complaint alleges the caller used an artificial voice, failed to disclose that fact at the start, and contacted a number on the Do Not Call Registry. After a short qualification pitch-offering a cash-out refinance to pay off high-interest debt or fund home improvements-the agent transferred the call to a human representative, which helped connect the outreach back to the lender.
The suit seeks class certification so other recipients of similar calls can be included. Separately, a related class action against medical marketing firm Altrua Healthshare over nondisclosed prerecorded calls reached a tentative settlement in February, signaling more plaintiffs are testing these claims.
How to use AI voice safely (and still hit targets)
- Flip the funnel: Collect consent via inbound lead gen (forms, SMS, web chat) and call only opted-in contacts.
- Lead scoring over blasting: Prioritize high-intent segments and reserve calls for those with explicit consent.
- Multichannel preview: Warm up with compliant email/SMS before any AI-assisted call attempt.
- Measure what matters: Transfers per 100 calls, complaint rate, opt-out rate, and human-takeover ratio.
Keep learning
Bottom line: AI can help you scale outreach, but consent, disclosure, and discipline come first. Get those right, and you'll grow without waking up to a class action.
Your membership also unlocks:
