AI Voice Calls and the TCPA: What Home Service Companies Need to Know

Your HVAC company just deployed an AI voice agent to follow up on estimates. It's booking appointments, freeing up your CSRs, and customers seem happy. Then a letter arrives from an attorney representing a class of consumers claiming you made thousands of illegal robocalls.

This scenario is playing out across home services as companies adopt AI voice technology without understanding the regulatory landscape. In February 2024, the FCC confirmed that AI-generated voices qualify as "artificial voices" under the Telephone Consumer Protection Act (TCPA). That ruling means every AI voice call your business makes is subject to the same federal regulations that govern robocalls and autodialed calls.

The good news: TCPA compliance isn't complicated once you understand the rules. This article breaks down what the law requires, how it applies to common home service use cases like appointment reminders and estimate follow-up, and what you need to do operationally to stay compliant.

Key takeaways

• AI voice calls trigger TCPA requirements. The FCC's 2024 ruling means any call using AI-generated voice technology is considered an "artificial voice" under federal law, requiring proper consent and disclosures regardless of the underlying technology.
• Marketing calls require written consent; informational calls require oral consent. Following up on an estimate with a sales pitch requires prior express written consent, while appointment reminders only need prior express consent. Misclassifying a marketing call as informational creates liability.
• Penalties reach $500 to $1,500 per call with no cap. A 1,000-call campaign without proper consent can create $500,000 to $1,500,000 in potential liability, making compliance essential before scaling AI voice outreach.
• Opt-out rules tightened in 2025. Businesses must honor revocation requests within 10 business days and recognize natural-language opt-out phrases like "stop calling me" or "leave me alone."
• Platform features support compliance, but configuration is your responsibility. AI communications platforms can provide consent verification and opt-out tracking, but you must capture proper consent and classify your campaigns correctly.

What the TCPA means for AI voice calls

The Telephone Consumer Protection Act, passed in 1991, restricts how businesses can use automated calling technology to contact consumers. For decades, the law focused on traditional robocalls and autodialed calls. The FCC's February 2024 declaratory ruling expanded that scope to include AI.

Why AI voices fall under TCPA scrutiny

The FCC's ruling in FCC 24-17 established that AI-generated voices meet the definition of "artificial or prerecorded voice" under the TCPA. This applies whether your AI agent is:

• Reading from a pre-written script
• Generating responses conversationally using large language models
• Using voice cloning or text-to-speech technology

The technology behind the voice doesn't matter. If a live human isn't speaking in real time, the call is subject to TCPA requirements. This means your AI voice agent for lead follow-up is treated the same as a pre-recorded robocall under federal law.

{{in-line-cta}}

How this applies to home service businesses

Home service companies commonly use AI voice technology for several customer communication scenarios:

• Speed-to-lead response: Calling back web leads within minutes of form submission
• Appointment reminders: Confirming tomorrow's HVAC tune-up or plumbing appointment
• Estimate follow-up: Reconnecting with homeowners who received a quote but haven't scheduled
• Customer reactivation: Reaching out to past customers due for maintenance or annual service
• Service updates: Notifying customers about technician arrival times or schedule changes

Each of these use cases falls under TCPA regulation. The specific consent requirements depend on whether the call is classified as informational or marketing.

Two types of consent: informational vs. marketing calls

The TCPA creates a two-tier consent framework. Understanding which tier applies to each type of call you make is the foundation of compliance.

Prior express consent for informational calls

Informational calls include appointment reminders, service updates, scheduling confirmations, and other communications that don't include a sales pitch or promotional content. These calls require prior express consent, which can be:

• Oral (the customer gave you their phone number and agreed to be called)
• Written (but a signed agreement isn't required)

A homeowner who fills out your website form and provides their phone number has typically given prior express consent to receive informational calls at that number, provided your form states that calls may be made.

Example: Your AI agent calls to confirm a roofing inspection scheduled for Thursday at 10 AM. The call reminds the homeowner to clear access to the attic and provides your office number for rescheduling. This is informational.

Prior express written consent for marketing calls

Marketing calls include any communication that promotes a product or service, encourages a purchase, or contains promotional language. These calls require prior express written consent, meaning:

• A signed written agreement (electronic signatures are acceptable)
• Clear disclosure that the consumer authorizes calls from your specific business
• Disclosure that calls may use automated or AI-generated voice technology
• The specific phone number being consented to
• Statement that consent is not required to purchase goods or services

Example: Your AI agent calls a homeowner who received an estimate last week. The script says, "I'm following up on your window replacement quote. We're currently offering 15% off installation if you schedule this month." This is marketing and requires written consent.

The critical distinction that creates liability

The most common TCPA violation pattern in home services is misclassifying marketing calls as informational. This happens when:

• An "appointment reminder" includes promotional language ("Ask about our spring savings!")
• An "estimate follow-up" uses sales language to encourage booking
• A "service update" includes upsell offers

One sentence of promotional content transforms an informational call into a marketing call. If you only have oral consent, you're now making an illegal marketing call.

How to stay compliant: Review every AI script for promotional language. If the call includes any encouragement to purchase, discount offers, or sales pitches, classify it as marketing and ensure written consent exists before the call is placed.

Required disclosures and opt-out mechanisms

Beyond consent, the TCPA requires specific disclosures during AI voice calls and mechanisms for consumers to opt out.

What your AI agent must say

Within the first few seconds of every call, your AI agent should provide:

1. Caller identification: "This is [Agent Name] calling from [Your Business Name]"
2. Callback number: A phone number the consumer can use to reach your business
3. Opt-out mechanism: Within two seconds of the initial message, offer a way to stop future calls

While not yet federally required, an AI disclosure ("This call uses automated voice technology") is recommended and may become mandatory. Some states already require disclosure of AI use in certain contexts.

Example opening script for an appointment reminder:

"Hi, this is Alex calling from Metro Plumbing to confirm your appointment tomorrow at 2 PM. You can reach us at 555-0123. To stop these reminders, say 'stop' or press 9."

Opt-out requirements after April 2025

The FCC tightened opt-out rules effective April 2025. Your business must now:

• Honor revocation by any reasonable means. Consumers can opt out by saying "stop," "quit," "end," "cancel," "unsubscribe," or even informal phrases like "leave me alone" or "don't call me again."
• Process opt-outs within 10 business days. This is down from 30 days under previous rules.
• Send only one confirmation message within 5 minutes of revocation, and it cannot contain marketing content.

As of November 2026, callers may designate an exclusive opt-out method for informational calls (such as a specific website or phone number). However, marketing opt-outs still apply across all marketing communications from your business.

Operational implications

Your AI voice agent must be configured to:

• Recognize natural-language opt-out requests mid-conversation
• Immediately stop the call when an opt-out is detected
• Update your internal Do-Not-Call list in real time
• Sync the suppression across all outbound campaigns, including voice, text, and email
• Retain opt-out records for at least five years

If your AI only recognizes the keyword "STOP," you're exposed. Train it to detect opt-out intent even when customers phrase requests informally.

The Do Not Call Registry and AI calling

The National Do Not Call Registry adds another compliance layer for marketing calls.

When registry scrubbing is required

If you're making marketing calls, you must scrub your call lists against the National DNC Registry. The registry data must be no more than 31 days old. You must also maintain your own internal Do-Not-Call list for consumers who have specifically requested not to receive calls from your business.

When the registry doesn't apply

The DNC Registry generally doesn't restrict calls to consumers who have given prior express consent. If a homeowner filled out your contact form requesting an estimate, you can follow up even if they're on the national registry.

However, if you're purchasing leads from aggregators and making unsolicited marketing calls, DNC scrubbing is mandatory. The safest practice is real-time scrubbing at call initiation, since daily registry updates can create gaps in batch scrubbing.

Internal Do-Not-Call list requirements

Beyond the national registry, you must maintain an internal list of consumers who have asked your specific business not to call them. This list must be:

• Updated immediately when opt-outs are received
• Retained for at least five years
• Applied to all marketing campaigns before calls are placed

Penalties for non-compliance

TCPA violations carry significant financial risk, and enforcement has accelerated alongside AI adoption.

Statutory damages

The TCPA allows consumers to recover:

• $500 per violation for negligent violations
• $1,500 per violation for willful or knowing violations

There is no cap on total damages. This means a 1,000-call AI campaign without proper consent creates $500,000 to $1,500,000 in potential exposure. A 10,000-call lead nurture campaign could result in $5 million to $15 million in liability.

Class action risk

Most TCPA cases are brought as class actions, aggregating claims from many consumers who received similar calls. TCPA class action filings increased 95% in the first half of 2025 compared to 2024, with AI voice calling emerging as a growing target for plaintiff attorneys.

Common violation patterns in home services

• Consent gaps on transfers: Your AI agent qualifies a lead and transfers them to a salesperson or financing partner. If the receiving party didn't obtain proper consent, both the AI platform operator and the transfer recipient face liability.
• Marketing disguised as informational: "Hi, this is a reminder about your estimate... and by the way, we're running a 20% off promotion this week!" This call crossed into marketing but likely only had informational consent.
• Opt-out failures: A customer says "stop calling me" during a conversation, but the AI doesn't recognize it or doesn't sync to the suppression list. The next campaign dials them again.

State-level requirements beyond federal TCPA

Federal TCPA compliance is the baseline, but many states have their own telemarketing restrictions that go further.

State mini-TCPA laws

State laws may impose:

• Narrower calling windows (some states restrict calls before 9 AM or after 8 PM)
• Additional consent requirements
• Stricter penalties
• Industry-specific rules

California, for example, has disclosure requirements for AI use in certain contexts. Illinois' Biometric Information Privacy Act (BIPA) may apply if voice data is used for identification purposes. Colorado's AI Deployment in Material Transactions framework imposes disclosure and record-retention rules when AI influences significant decisions like qualifying leads for financing.

Two-party consent states for call recording

If your AI voice calls are recorded for quality or compliance purposes, you must disclose the recording in two-party consent states, including California, Connecticut, Florida, Illinois, Maryland, Massachusetts, Montana, Nevada, New Hampshire, Pennsylvania, and Washington.

Practical guidance for multi-state operations

If you operate in multiple states:

1. Identify every state you call into
2. Research that state's mini-TCPA and call recording laws
3. Apply the strictest standard across all campaigns (easier than maintaining state-by-state scripts)
4. Document your research and policy decisions

For businesses calling into many states, working with telecom or TCPA counsel to audit your program before launch is often worthwhile.

Implementation checklist for home service companies

Use this checklist to build or audit your TCPA compliance program for AI voice calls.

Audit your current practices

Classify every AI voice use case:

• List every scenario where AI makes outbound calls
• Determine whether each is informational or marketing
• Document the business purpose and phone numbers called

Review consent records:

• Do you have written consent for marketing calls?
• Does consent mention AI or automated calls?
• Is consent tied to specific phone numbers?
• Where is consent stored (CRM, form submissions)?
• Can you produce consent records within 24 hours if needed?

Check existing scripts:

• Do AI agents identify your business by name?
• Do they provide a callback number?
• Do they offer an opt-out mechanism within 2 seconds?
• Does any "informational" script include promotional language?

Fix consent gaps

Update web forms and lead capture:

• Add TCPA-compliant consent language to all forms
• Example: "By providing your phone number, you consent to receive calls and texts from [Business Name], including those made using automated or AI-generated voice technology, at the number you provide. Consent is not required to purchase services."
• Use e-signature capture for marketing campaigns
• Store consent with timestamp, IP address, and exact form language

Configure CRM integration:

• Ensure consent status syncs to your AI platform
• Set rules preventing outbound AI calls without consent verification
• Map consent fields to campaign classifications (informational vs. marketing)

Update AI agent scripts

Required elements:

• Opening identification within first 5 seconds
• Callback number
• Opt-out mechanism within 2 seconds of initial message
• Natural-language opt-out recognition throughout the call

Test before scaling:

• Run scripts through compliance review
• Test with live calls in low-volume scenarios
• Verify opt-out detection works with various phrasings

Establish ongoing safeguards

Call logging and audit trail:

• Log every call with number, time, call type, script version, and outcome
• Store consent record ID with each call attempt
• Retain call recordings and transcripts (with proper two-party consent disclosures)
• Keep all records for at least five years

Do-Not-Call list management:

• Maintain internal DNC list
• Scrub marketing lists against National DNC Registry (31-day max age)
• Sync opt-outs across all channels
• Process opt-outs within 10 business days

Monitoring and training:

• Review sample AI calls weekly for script compliance
• Monitor opt-out volume for spikes that may indicate problems
• Train CSRs on TCPA rules for handling escalations and transfers
• Track regulatory changes quarterly

How Hatch supports TCPA-compliant customer communications

When you're managing customer conversations across the full journey, from the first inquiry through booking, follow-up, and ongoing care, compliance becomes an operational challenge across multiple touchpoints. Hatch is an AI communications platform built for service businesses that includes compliance infrastructure alongside its customer communication capabilities.

Built-in compliance support

Hatch's platform includes SOC 2 verification and built-in TCPA/10DLC support through the onboarding process. This means consent language templates for web forms, "customer must text first" enforcement for certain campaigns, and native CRM integration that can verify consent status before AI initiates outreach.

The platform connects with systems service businesses already use, including ServiceTitan, LeadPerfection, improveit 360, and AccuLynx through Data Bridge. This integration allows consent status to travel with the customer record, reducing the gap between where consent is captured and where calls are initiated.

Operational compliance features

Voice AI and Messaging AI include natural opt-out recognition, detecting phrases like "stop," "quit," and "leave me alone" to update suppression lists and end outreach. When customers do want to continue the conversation, warm transfers preserve context so the handoff to your team doesn't create consent gaps.

Journey Builder allows you to design automated customer communication flows with time-zone-aware scheduling and business hours enforcement. Command Center provides full conversation review and export for compliance documentation, plus performance reporting that can flag opt-out rate spikes early.

What remains your responsibility

Hatch provides compliance infrastructure, but several elements require your attention:

• Consent form language: Hatch provides templates, but legal counsel should review your specific consent language
• Campaign classification: You determine whether each campaign is marketing or informational
• National DNC Registry scrubbing: You or your CRM must handle registry scrubbing for marketing calls
• State law compliance: State-specific requirements are your responsibility to monitor and implement

The platform helps you manage compliant customer conversations at scale. The compliance decisions remain with your business.

Getting started with compliant AI voice calls

TCPA compliance for AI voice calls comes down to three fundamentals: proper consent, required disclosures, and functional opt-out handling. Get these right, and AI voice technology becomes a legitimate tool for responding faster to leads, confirming appointments, following up on estimates, and staying connected with customers.

Start by auditing your current AI calling practices against the consent framework. Classify each use case, review your consent records, and check your scripts for the required elements. Fix gaps before scaling. Consider working with telecom or TCPA counsel if you operate in multiple states or handle high call volumes.

The businesses that adopt AI voice technology with compliance built in from the start will have an advantage over those scrambling to retrofit compliance after receiving demand letters. The rules are clear. The question is whether you implement them before or after they become expensive.