Why the Autodialer Definition Trips Up Businesses

If you run outbound calling campaigns, there's a question you've probably asked: does my dialing setup count as an autodialer? The answer shapes which consent rules apply to every call you make and how much legal exposure you carry if something goes wrong.

It's not a small question. The TCPA allows $500 to $1,500 in statutory damages per call, and lawsuits can look back four years for violations. A single misclassified campaign can turn into six figures of exposure before you've finished reading the demand letter.

Before 2021, the answer depended on where you got sued. Six federal circuits split on what "capacity" meant when describing a dialing system. The Third, Seventh, and Eleventh Circuits read the definition narrowly. The Second, Sixth, and Ninth Circuits read it broadly enough that almost any modern phone could qualify.

That split created three kinds of confusion that hung over every compliance team at the time:

• The same dialing software could be an autodialer in New York and not one in Pennsylvania
• A system that merely *could* be modified to use a random number generator might count as an autodialer, even if it never did
• Consent obligations shifted based on geography rather than on how the technology actually worked

The Supreme Court resolved the split in Facebook, Inc. v. Duguid by applying the series-qualifier canon: the phrase "using a random or sequential number generator" modifies both "store" and "produce." Justice Sotomayor, writing for a unanimous Court, warned that the broader reading "would take a chainsaw to these nuanced problems when Congress meant to use a scalpel."

Here's what that means in practice. A system that stores numbers in a database and dials them from a fixed list is not an autodialer. It doesn't matter how fast the system dials or how many numbers it holds. What matters is whether the numbers come from a random or sequential generator.

That distinction matters more than most people realize, because it changes your consent tier. Autodialed marketing calls require prior express written consent. Non-autodialed, manually dialed calls fall under a lighter standard: prior express invitation or permission.

At My AI Call Center, we dial from approved, permissioned, or reviewed contact lists only. We never use random or sequential number generators to produce numbers. We check list source and consent records before any campaign launches, and we tell you plainly if a list won't support the campaign before you spend anything.

One important caveat: narrowing the autodialer definition didn't erase other TCPA rules. Artificial and prerecorded voice restrictions remain fully intact and carry their own consent requirements. If your system uses AI-generated voices, those are treated as artificial voices under the statute, so prior express consent is required regardless of how the dialer classifies.

The bottom line: stop guessing about "capacity" and look at the actual mechanism. If your numbers come from a list, not a generator, you're outside the autodialer definition. Document that fact, keep your consent records tight, and you've removed the single largest source of TCPA uncertainty from your operation.

The Facebook v. Duguid Standard: Random or Sequential Generator Required

In 2021, the Supreme Court finally answered the question that had divided federal courts for years. In Facebook, Inc. v. Duguid, a unanimous Court held that an autodialer must have the capacity to store or produce telephone numbers using a random or sequential number generator. Full stop.

Before that ruling, six federal circuits were split on the definition. The Third, Seventh, and Eleventh Circuits read the statute narrowly, while the Second, Sixth, and Ninth Circuits took a broad view that could sweep in ordinary list-dialing systems. The Supreme Court ended the fight by applying the "series-qualifier canon," determining that the phrase "using a random or sequential number generator" modifies both verbs in the statute: "store" and "produce." A system that merely dials numbers saved in a database doesn't qualify, because neither verb is satisfied without the generator.

Justice Sotomayor, writing for the Court, explained why the narrower reading matters. Expanding the definition to cover any system that dials stored numbers "would take a chainsaw to these nuanced problems when Congress meant to use a scalpel." The TCPA, she wrote, targets a unique type of telemarketing equipment, not every piece of technology that can place a call automatically. The D.C. Circuit had already warned that broader readings would create an "eye-popping sweep" that could bring every smartphone under the statute.

The practical takeaway for any business running outbound calls is straightforward:

• If your system dials from a fixed, pre-approved list without a random or sequential number generator, it generally does not meet the federal definition of an autodialer.
• If the system does use such a generator to store or produce numbers, it qualifies, and the strictest consent requirements apply.
• The FCC's current position focuses on what your equipment can do right now, not what it could theoretically be modified to do.
• Narrowing the autodialer definition does not touch the TCPA's separate rules on artificial or prerecorded voices, which remain fully intact.

That last point matters more than most businesses realize. The stakes here are real: TCPA violations carry statutory damages of $500 to $1,500 per call, and willful violations can reach $1,500 each, with a four-year lookback window for lawsuits. Whether your dialer meets the Duguid standard often determines which consent tier you fall into.

This is why list discipline is central to how we work at My AI Call Center. Campaigns run against approved, permissioned, or reviewed contact lists only, and we check list source and consent records before anything launches. Because the dialing draws from a structured list rather than generated numbers, the Duguid standard is the benchmark we build around. It's also why we tell clients plainly when a list won't support a campaign, before they spend anything. Knowing what qualifies as an autodialer is the first step in knowing where your liability actually sits.

The Catch: Narrowing the Dialer Definition Didn't Narrow the Rules

Here's the nuance most businesses miss: winning the argument that your dialer isn't an autodialer doesn't put you back to a consent-free world. The TCPA's rules on artificial and prerecorded voices survive completely untouched, and those rules apply to you the moment your calls use anything other than a live human voice.

The Supreme Court's 2021 Facebook v. Duguid ruling narrowed the definition of the dialing device, requiring a random or sequential number generator. It did nothing to soften the separate prohibition on artificial or prerecorded voice calls, which remains fully intact with its own consent requirements. That distinction matters enormously if your calls use AI-generated voices, which are treated as artificial voices under the TCPA and require prior express consent regardless of how the numbers got dialed.

In practice, this creates two separate compliance questions for every campaign:

• Does the dialing system store or produce numbers with a random or sequential generator? If yes, you're using an autodialer and face the strictest consent tier.
• Is the voice artificial or prerecorded? If yes, prior express consent is required even if the dialer itself falls outside the autodialer definition.
• Do your consent records actually prove permission? This is where campaigns fail, not at the technology layer.

The consent documentation question got a partial reset in 2026, when the Fifth Circuit ruled in Bradford v. Sovereign Pest Control that the TCPA does not require prior express written consent for automated calls, accepting oral consent as valid. That's a genuine win for businesses that take permission by phone. But the same analysis comes with a warning worth heeding: oral consent must be clear, direct, and unequivocal, and it should be carefully documented and independently verifiable to withstand future scrutiny.

"Independently verifiable" is the operative phrase. A spreadsheet note saying "customer agreed on 3/14" won't survive a four-year lookback in litigation, which is how far TCPA claims can reach. Timestamped call recordings and logs tied to the specific number called are the standard that holds up.

This is why consent review sits at the front of every campaign we run at My AI Call Center. We check list source and consent records before anything launches, flag bought lists that lack clear permission records, and in most cases decline them outright. It's slower than just dialing, and it's the only way we know a campaign is defensible before you spend anything.

The takeaway: narrowing the dialer definition narrowed one risk. Voice technology rules and consent documentation carry the rest, and they're where the real exposure lives.

How to Check Your Own Setup (and What to Ask Any Calling Vendor)

Knowing the legal definition of an autodialer is one thing. Confirming your own dialing setup actually meets that definition is where real compliance lives.

Start with the dialing mechanism itself. Under the Supreme Court's Facebook v. Duguid ruling, a system only qualifies as an autodialer if it can store or produce numbers using a random or sequential number generator, so equipment that merely dials from a fixed list falls outside the definition (Wiley's analysis). Ask your vendor directly: does the system generate or store numbers randomly or sequentially, or does it dial from a static uploaded list? Get the answer in writing, because the FCC's current position focuses on what equipment can do right now, not what it could be modified to do (ActiveProspect explains).

Next, audit your consent records. TCPA lawsuits can look back four years, and statutory damages run $500 to $1,500 per call, so you need documentation for the entire limitations window (ActiveProspect notes). And if you rely on oral consent, remember that even courts accepting it expect records that are carefully documented and independently verifiable.

Your opt-out handling deserves equal attention. Since April 11, 2025, the rule requires honoring opt-out requests within 10 business days, but treat that as a ceiling, not a target. Suppressing a contact the moment they say "stop" removes any risk of a timing dispute and keeps your suppression list clean.

Here's a practical audit list for any dialing setup, in-house or vendor-run:

• Confirm in writing that no random or sequential number generator stores or produces the numbers being dialed.
• Document list source and consent records for every contact, covering all four years of the limitations window.
• Honor opt-outs immediately rather than waiting out the 10-business-day deadline.
• For DNC-listed contacts, verify the relationship existed within the past 18 months before relying on the Established Business Relationship exemption.

That EBR point trips up a lot of teams. The exemption for DNC-registered numbers only applies if a genuine business relationship existed within the last 18 months, and the burden of proving that relationship falls on you. A dormant contact from three years ago doesn't qualify, no matter how warm the relationship once was.

One more thing: even a dialer that avoids autodialer status doesn't get you off the hook entirely. The Duguid ruling left restrictions on artificial and prerecorded voices fully intact, so if you're using AI-generated voices, consent requirements still apply regardless of how the numbers get dialed.

This is why list discipline, not dialing technology, is the real compliance strategy. It's the approach My AI Call Center takes: every campaign runs only against approved, permissioned, or reviewed lists, with list source and consent records checked before launch, and bought lists without clear permission records flagged or declined. Whether you run campaigns yourself or hire a vendor, the questions above will tell you quickly whether your setup can withstand scrutiny.