Back to all articles

Are Automated Calls Compliant? The Real Rules

Relay by Cactus AI

Are Automated Calls Compliant? The Real Rules

A homeowner fills out a form for a roof estimate at 7:15 p.m. Your team is wrapping up for the day. An automated caller reaches them in two minutes, qualifies the job, and transfers a real prospect to the on-call rep. That can be a good use of automation. It can also create a compliance problem if the lead, call type, consent, or follow-up process is wrong.

So, are automated calls compliant? They can be. But compliance is not a switch you turn on by choosing a reputable phone platform or adding a disclosure to a script. It comes down to who you call, why you call, what technology delivers the message, where the person lives, and how you handle consent and opt-outs.

For a business that lives on booked jobs and qualified transfers, the goal is not to avoid calling. The goal is to build a calling process that creates revenue without creating complaints, blocked numbers, or legal exposure.

Are automated calls compliant under federal rules?

Federal law puts the biggest restrictions on certain calls and texts made using an artificial voice, prerecorded voice, or automated dialing technology. The Telephone Consumer Protection Act, usually called the TCPA, is the rule most business owners hear about first.

The practical takeaway is straightforward: marketing calls using a prerecorded or artificial voice to a cell phone generally require prior express written consent. That consent needs to be clear, tied to the seller, and obtained before the call. A purchased list, a business card from three years ago, or a vague agreement to receive "updates" may not be enough for an automated marketing call.

The Federal Communications Commission has also said that AI-generated voices fall within the TCPA's restrictions on artificial or prerecorded voices. If an AI voice agent is making an outbound marketing call, treat it as a regulated artificial voice campaign. Do not assume the call is treated like a live employee simply because it can hold a natural conversation.

There are exceptions and different standards for informational, transactional, emergency, and debt-related calls. A dentist confirming tomorrow's appointment is not the same as a sales agent calling a cold list about kitchen remodeling. But the line can get blurry fast. A reminder call that pivots into a promotional offer may no longer look purely informational.

This is why "we only call people who gave us their number" is not a compliance strategy. You need to know what they agreed to receive and preserve the proof.

Consent is the part that breaks most campaigns

For outbound marketing, start with the lead source. A lead from your own website may be workable if the form includes the right consent language and records the person's agreement. A lead from a third-party marketplace may be workable, but only if that source can show valid consent for your specific company and the type of calls you plan to make.

That last part matters. Consent to hear from a lead vendor or a broad group of "marketing partners" does not automatically mean your business has permission to place an AI voice call. The exact wording, the seller named in the disclosure, the date and time, the phone number entered, and the page version all matter when a complaint arrives months later.

A practical consent record should capture at least four things:

  • The phone number and lead source
  • The consent language shown to the consumer
  • The date, time, and method of consent
  • Evidence that the consumer submitted or otherwise agreed to it

Keep the record connected to the contact in your CRM. If you cannot quickly answer, "Where did this person opt in, and what did they agree to?" do not run that record through an automated marketing campaign.

There is a real trade-off here. Tightening your consent standards may shrink the number of leads you can call. It also improves answer quality, reduces complaints, and keeps your best phone numbers from being tagged as spam. A smaller list of people who actually expect to hear from you often produces more booked work than a giant list worked recklessly.

Cold calling needs a different playbook

Cold outreach is where many operators get burned. A business may be allowed to make some manually dialed live-agent calls to cold prospects, subject to do-not-call rules and other requirements. That does not mean an AI voice agent can call the same list with the same message.

If your outbound system uses an artificial or prerecorded voice, cold consumer lists are generally a poor fit for marketing. The risk is high because you likely do not have the written consent required for that call type.

That does not make outbound automation useless. It changes where it belongs. Use it on leads who have requested contact and consented, on existing customers where the call is genuinely service-related and legally permitted, or in workflows your counsel has reviewed. For cold acquisition, a live human calling process may have a different compliance analysis, but it still requires discipline around do-not-call rules, scripts, and state laws.

Business-to-business calls can also have different rules than consumer calls. Still, do not treat a business number as a free pass. Many owners use cell phones for work, state laws may apply, and a poorly targeted campaign can hurt your reputation even when it is technically allowed.

Do-not-call rules are operational, not paperwork

Compliance fails in the day-to-day details. Someone asks not to be called again. A rep writes a note but does not update the dialing list. The number gets called next week by a different campaign. One mistake becomes the kind of experience people report, post about, and remember.

Every outbound operation needs an internal do-not-call process that updates immediately. When someone says "stop," "don't call me," or any clear variation, honor it. Do not make them use a magic phrase. Do not transfer them around. Do not argue about whether they filled out a form.

Your calling system should suppress that number across campaigns, not just in the one list where the request happened. Train everyone who touches calls, including office staff and sales reps, on what an opt-out sounds like and what to do next.

You also need to screen applicable lists such as the National Do Not Call Registry and maintain your own company-specific do-not-call list. Established business relationships and inquiries can affect the analysis, but they are not permanent permission slips. Timing, the type of call, and state requirements still matter.

Calling hours, identification, and state law still count

A compliant lead can become a bad call if it lands at the wrong time. Federal rules generally limit telemarketing calls to a consumer's local hours, commonly 8 a.m. to 9 p.m. But states can set tighter rules. If your business serves multiple states, use the recipient's local time and follow the stricter rule when rules conflict.

Your caller should clearly identify the business and provide a way for the person to reach you. Hiding behind a vague brand name, a spoofed local number, or an unclear opening may lift short-term pickup rates, but it creates trust and compliance problems. Caller ID information should be accurate, and your business should be easy to identify.

State law is where a simple nationwide campaign gets complicated. Some states have their own telemarketing registration requirements, mini-TCPA laws, disclosure requirements, restricted calling windows, or private rights of action. Rules also change. A workflow that was reviewed last year may need another look after an FCC order, state law update, or change in how your team uses AI voices.

Build compliance into the call flow before launch

The fastest way to create risk is to treat compliance as a legal review at the end. Build it into how leads enter, move, and leave your system.

First, separate your lists. Mark records by source, consent status, state, purpose of the call, and whether an AI voice is permitted. Do not let a general lead list become a mixed bag of website inquiries, old customers, bought data, and unknown records.

Second, write a plain-language opening. The agent should state the business name, the reason for the call, and a clear route to opt out. If a prospect asks whether they are speaking to AI, do not dodge the question. A straightforward answer is better for trust and avoids a conversation that feels deceptive.

Third, test the handoff. If an AI agent qualifies a lead and warm-transfers them to a closer, make sure the closer sees the lead source, consent status, and notes from the call. A good transfer should save time, not force the customer to repeat themselves or put your team in the dark.

Finally, audit the operation. Review call recordings where permitted, opt-out logs, number reputation, complaint patterns, and consent records. A managed service such as Relay by Cactus AI should help keep these operational pieces organized, but the business making the calls still owns the campaign decisions and the legal risk.

Get a real answer for your exact campaign

No article can tell you whether every call your company makes is legal. The right answer depends on your state footprint, the lead source, whether the message is marketing or service, the phone number called, the voice used, and the rules in effect when you launch.

Have qualified counsel review your consent language, scripts, list sources, and call flows before you scale. Then give the person running the campaign simple rules they can follow without guessing.

The businesses that win with phone automation are not the ones that call the most numbers. They are the ones that answer fast, call people who expect to hear from them, honor a no without friction, and turn good conversations into booked jobs.