How courts define human intervention in dialing TCPA cases

Last updated 2026-07-11

What does the TCPA actually say about automated dialing?

The statute's definition is where everything starts. 47 U.S.C. § 227(a)(1) defines an automatic telephone dialing system as "equipment which has the capacity to store or produce telephone numbers to be called, using a random or sequential number generator; and to dial such numbers." ^[1] That's the full text. Every argument courts have about ATDS status flows from those words.

Notice the statute never mentions human intervention. Not once. The phrase entered the doctrine through FCC rulemaking and court interpretation, not through Congress. That gap is exactly why there's so much litigation. Courts have had to build their own tests to figure out what the statute means on a real call floor.

The FCC tried to fill that gap through decades of declaratory rulings. In its 2003 and 2008 orders, the agency took the position that equipment counts as an ATDS if it has the "capacity" to function as one, even when that capacity sits dormant. ^[2] That broad reading drove an explosion of suits. The agency also said systems requiring human intervention to dial don't qualify, but it never pinned down how much intervention matters or what kind.

How did Facebook v. Duguid change the ATDS definition?

Facebook, Inc. v. Duguid, 592 U.S. 395 (2021), is the biggest TCPA ruling since the statute passed. The Supreme Court resolved a circuit split and held that a device is an ATDS only if it uses a random or sequential number generator to store or produce numbers. ^[3] A system that just dials from a stored list, with no random or sequential generation, is not an ATDS under the statute.

Duguid himself got text messages from Facebook's security notification system, which pulled his number from a database of registered users and sent alerts when someone tried to log into his account. The Court ruled Facebook's system wasn't an ATDS because it didn't use a random or sequential number generator. The dialing was automated. The number source wasn't. That distinction decided the case.

For outbound sales teams, Duguid was mostly good news. Load a list of opted-in prospects, let a predictive dialer call them, and that CRM-to-dialer workflow looks much safer post-Duguid than it did before. But "much safer" is not "completely safe." Human intervention still matters, because defendants raise it as a backup defense, and some plaintiffs argue that modern dialers with certain queue-management features do something functionally equivalent to random or sequential generation.

What is the human intervention test and where did it come from?

The human intervention doctrine started with the FCC's 1992 rules implementing the TCPA and got refined in later orders. The core idea is simple. If a person must take a meaningful action to cause each individual call to be placed, the system doesn't function as an ATDS for that call.

The FCC's 2003 ruling stated that autodialers requiring human intervention to operate "each call" are not ATDSs. ^[2] Courts picked that up and started asking a harder question: how much intervention, of what kind, at what point in the process?

The leading appellate statement comes from Ninth Circuit cases interpreting the FCC's guidance. Marks v. Crunch San Diego, 904 F.3d 1041 (9th Cir. 2018), read the statute broadly before Duguid and held that even a system dialing from a stored list could be an ATDS. ^[5] Duguid largely displaced that logic, but its discussion of human intervention survives in lower court analyses.

Post-Duguid, human intervention arguments show up in one of two patterns. A defendant argues its system requires so much manual action that it isn't really automated at all, regardless of whether it uses random or sequential generation. Or a plaintiff argues that even a system that looks manual is really running the show, and the human clicks are just theater. Courts have gone both ways on both.

How much human involvement actually removes ATDS status?

This is where courts genuinely disagree, and you should not expect a clean answer.

The clearest cases sit at the extremes. An agent who physically dials a number by pressing each digit on a phone keypad is obviously not using an ATDS. A system that pulls numbers from a randomized list, queues them, and dials with no human in the loop is obviously an ATDS if it also uses random or sequential generation. The hard cases live in the middle.

Courts have found enough human intervention (and thus no ATDS) where a human must individually select each contact and click a send or call button before the system acts, where the system can't start any outreach without an affirmative per-contact decision by an agent, or where the platform works as a click-to-call tool and the agent controls timing while the system just does the mechanical dialing. ^[4]

Courts have found the intervention insufficient (ATDS status survives) where agents click a single button to launch a campaign and the system then works through hundreds of numbers on its own, where a supervisor approves a batch upload and the dialer fires off calls with no further input, or where the human action is upstream of the dialing itself, like uploading a list, while the actual calling is fully automated. ^[5]

The honest takeaway: courts look at what happens at the moment of each individual call, not at the campaign setup. Clicking "upload contacts" once and then having the system call 5,000 people in sequence is not human intervention under most courts' analyses.

Which circuits have the strictest view of human intervention?

Before Duguid, the Ninth Circuit had the broadest ATDS definition and, correspondingly, the strongest view that human intervention mattered. The Eleventh Circuit read ATDS more narrowly but was inconsistent on what intervention requires. The Second Circuit followed Duguid's logic closely once the Supreme Court ruled.

Post-Duguid, the circuit landscape on human intervention specifically is still forming. The Eleventh Circuit in Glasser v. Hilton Grand Vacations Co., 948 F.3d 1301 (11th Cir. 2020), decided before Duguid, held that a dialing system requiring a human agent to manually dial each number, even with software help, was not an ATDS. ^[4] That case still gets cited for the proposition that genuine per-call human control defeats ATDS classification.

The Ninth Circuit post-Duguid has been careful. In several district court decisions within the circuit, judges have held that systems where agents must individually approve each outbound connection pass the human intervention test, while systems where agents just set parameters and let the dialer run do not.

The difference between circuits matters most if you're choosing where to incorporate or do business. Ninth Circuit litigation is common for California companies. Eleventh Circuit cases come out of Florida, Georgia, and Alabama, where outbound call centers are numerous. Neither circuit is a safe harbor. They just weigh the factors a little differently.

For teams doing cold calling at scale, the safest move post-Duguid is to document exactly what a human agent does before each call fires, not before each campaign launches.

Does a preview dialer count as having sufficient human intervention?

Preview dialers are the tool defendants cite most in human intervention arguments, and courts have split on them.

In a preview dialer workflow, an agent sees a contact record on screen, reviews it, then clicks a button to start the call. The system does the mechanical dialing, but a human decided to call that specific person at that specific moment. Many courts have found this enough. The agent's click is a per-contact, pre-dial decision, and the system can't call anyone the agent hasn't individually approved.

Power dialers and predictive dialers are a different animal. In a power dialer, the system works through a list automatically and connects agents only when a call is answered. Agents don't choose individual contacts. Courts have generally been colder toward human intervention arguments for power dialers, especially if the system can detect and leave voicemails with no agent action.

Predictive dialers, which call multiple numbers at once and drop or transfer based on answer detection, fare the worst under human intervention analysis. The whole point of a predictive dialer is to strip human timing decisions out of the dialing process. ^[7]

If you're evaluating your text message marketing or call stack, ask your vendor one specific question. Does a human take a per-contact action immediately before each call or message is sent, or does a human just configure and launch a campaign? That answer drives most of the legal analysis.

What evidence do courts look at to determine human intervention?

Courts don't take defendants' word for how their systems work. Plaintiff's counsel routinely requests system logs, software documentation, vendor contracts, and employee training materials during discovery. Here's what courts actually examine.

System architecture logs. Does the system record agent actions at the individual call or message level? If logs show campaign launches but not per-contact approvals, that's evidence against human intervention.

Vendor documentation. How does the vendor market the product? A sales deck calling the tool a "set it and forget it" autodialer is a bad fact for the human intervention defense.

Employee testimony. What do agents actually do during a calling shift? If agents testify they mostly wait for connected calls and handle conversations, that undercuts claims of per-contact human control.

Call timing patterns. Plaintiffs' expert witnesses often analyze call timing data statistically. Calls placed at machine-speed intervals, or calls placed to sequential numbers within a short window, suggest automation even if a human was nominally involved.

Dropped call rates. High abandoned call rates are a signature of predictive dialers. Courts and the FCC have noted this connection. ^[2]

One thing that genuinely helps defendants: contemporaneous written policies requiring agents to review contact records before dialing, paired with system logs that actually show those review steps. Courts are far more receptive to human intervention arguments when the evidence predates the lawsuit.

Does human intervention matter for text messages the same way it does for calls?

Yes, and the analysis is nearly identical. The TCPA covers text messages sent using an ATDS, and courts apply the same ATDS definition from 47 U.S.C. § 227(a)(1) to texts. ^[1] The human intervention doctrine applies to SMS and MMS campaigns exactly the way it applies to outbound calls.

The practical difference is that text platforms make batch sending easy. A user can upload a list, set a message template, and send to 50,000 contacts with one button click. Almost every court that has looked at mass text campaigns has found insufficient human intervention, because the human action (clicking send on a campaign) sits too far upstream from each individual message to count.

One-to-one texting tools, where an agent composes and sends each message individually to one contact at a time, look much better under the human intervention test. Some platforms market themselves as manual, agent-driven tools partly for this reason.

The FCC's 2023 order on robotexts flagged automated text campaigns as a priority enforcement area. ^[8] That signals the agency views most mass texting platforms as ATDS systems regardless of human intervention claims. Courts don't have to follow FCC enforcement priorities, but agencies can file amicus briefs, and agency interpretations still carry some weight in federal court.

LeadCompliant's free TCPA tools include a dialer-type checker that helps you categorize your platform's human intervention profile before a plaintiff's expert does it for you.

What are the biggest human intervention cases after Duguid?

Duguid narrowed the ATDS definition, so many post-2021 cases turn on the random-or-sequential-number-generator requirement rather than human intervention alone. But human intervention arguments still appear regularly, especially when defendants argue their systems required too much manual action to qualify even under the broader pre-Duguid standard.

Some notable post-Duguid district court patterns:

Case pattern	Human intervention finding	Dialer type
Per-contact click-to-dial, agent-controlled	Sufficient intervention, no ATDS	Preview dialer
Campaign launch, automated execution	Insufficient intervention, ATDS analysis proceeds	Power/predictive dialer
Batch SMS, single send button	Insufficient intervention	Mass text platform
Individual agent-composed texts, one at a time	Sufficient intervention	Manual SMS tool
Voicemail drop, no agent per call	Insufficient intervention	Ringless voicemail

The cash app TCPA class action settlement and similar high-profile cases show that ATDS status combined with lack of consent produces the largest exposure. Human intervention is rarely the only issue. It's usually argued alongside consent disputes.

The credit one TCPA settlement involved allegations about automated calling at scale and shows how fast per-call statutory damages (up to $500, or $1,500 for willful violations) pile up into class-action territory. ^[1]

What should outbound sales teams do right now based on how courts are ruling?

The steps aren't complicated. Most teams just skip them.

First, document your dialer's actual workflow at the per-call level. Write it down. What does an agent do before each call fires? If the answer is "nothing, the dialer handles it," you have an ATDS problem that human intervention arguments won't solve.

Second, if you're using a power dialer or predictive dialer, ask counsel whether your calling practices survive Duguid's random-or-sequential-generation test before you lean on human intervention as a fallback. Duguid helps a lot of defendants. It's not blanket immunity.

Third, get consent documentation in order regardless of ATDS status. Even a fully manual cold call to a cell phone triggers TCPA compliance if the call is telemarketing and the recipient is on the do not call list. Human intervention defeats ATDS classification. It doesn't defeat DNC obligations.

Fourth, if you use any form of mass texting, don't assume your vendor's marketing claims about being "human-driven" will survive discovery. Request system architecture documentation and have counsel review it.

Fifth, train agents on what their click means legally. If the human intervention defense depends on agents individually approving each contact, agents need to know that and actually do it, rather than clicking through a queue as fast as they can. Courts look at real workflows, not policy documents.

The mobile phone do not call list rules layer on top of all of this. Winning on ATDS status doesn't help you if you're calling cell numbers on the national DNC registry without prior express invitation or permission.

Are there FCC rules pending that could change the human intervention analysis?

The FCC has been busy. Its 2023 declaratory ruling on one-to-one consent tightened the consent side of TCPA enforcement significantly. ^[8] The agency has not, as of mid-2026, issued a new declaratory ruling that directly redefines what counts as sufficient human intervention post-Duguid. That doesn't mean it won't.

There's ongoing industry pressure on both sides. Some carrier and technology groups want the FCC to clarify that modern AI-assisted dialers with human oversight don't qualify as ATDSs. Consumer advocacy groups want the opposite. The FCC can issue interpretive rules and declaratory rulings that courts give Skidmore deference, meaning courts can follow FCC interpretations if they find them persuasive, even though they aren't strictly bound by them after Loper Bright. ^[9]

Here's the practical implication. If the FCC issues a new order on human intervention, expect a wave of litigation testing whether courts follow it. Build your compliance posture around the current statutory text and Duguid rather than betting on favorable agency guidance.

For teams tracking regulatory developments, the FCC publishes its rulemaking dockets online. TCPA matters live in the CG docket series, and the main proceeding is CG Docket 02-278. ^[10]

LeadCompliant publishes summaries of new FCC orders as they issue, with compliance checklists tied to each. The brand's free compliance kit covers dialer classification, consent documentation, and DNC scrubbing in one place.

Frequently asked questions

What is the human intervention test in TCPA cases?

The human intervention test asks whether a person takes a meaningful, per-contact action immediately before each call or text is placed. If an agent must individually decide to call or text each recipient and trigger that communication, most courts find sufficient human intervention and decline to classify the system as an ATDS. If a human only sets up a campaign and the system does the rest, courts typically find insufficient intervention.

Did Facebook v. Duguid eliminate the human intervention doctrine?

No. Duguid narrowed the ATDS definition by requiring random or sequential number generation, which cut how often ATDS status attaches in the first place. But human intervention remains a live defense. Defendants whose systems use random or sequential number generation still argue human oversight defeats ATDS classification, and courts continue to analyze what level of involvement counts.

Does clicking send on a mass text campaign count as human intervention?

Almost universally, no. Courts that have examined mass SMS campaigns have found that a single click launching messages to thousands of recipients is not per-contact human intervention. The human action is too far removed from each individual message. Manual, agent-composed, one-at-a-time texting is the only text workflow that reliably passes the human intervention test in most circuits.

Is a predictive dialer always an ATDS under current law?

Post-Duguid, a predictive dialer is an ATDS if it also uses random or sequential number generation to produce the numbers it dials. If it dials only from a preloaded list with no random or sequential generation, Duguid's holding suggests it may not be an ATDS by definition. But predictive dialers almost never have sufficient human intervention either, so both prongs of the analysis tend to go against defendants using them.

What dialer type is safest from a human intervention standpoint?

Preview dialers, where an agent reviews a contact record and clicks to start each individual call, have the strongest human intervention profile. The agent makes a per-contact decision before every call. Power dialers and predictive dialers require far less per-call human action and hold up poorly under the human intervention test. Click-to-dial tools built into CRM systems like Salesforce or HubSpot generally resemble preview dialers.

Can I win a TCPA case just by proving human intervention?

Probably not on its own. Human intervention defeats ATDS classification, which removes one basis for liability under 47 U.S.C. § 227(b). But TCPA claims can also arise from DNC registry violations under 47 U.S.C. § 227(c), which apply to any telephone solicitation call regardless of whether an ATDS was used. Winning the ATDS fight still leaves you exposed if you're calling numbers on the national DNC list without proper consent.

What evidence do courts use to evaluate human intervention claims?

Courts look at system logs showing per-contact agent actions, vendor marketing materials describing how the platform works, employee testimony about actual call-floor workflows, call timing data analyzed by expert witnesses, and dropped-call rates. Defendants who can produce contemporaneous logs showing per-contact agent approvals fare better than those who rely only on policy statements made after a lawsuit was filed.

Does the human intervention doctrine apply differently in the Ninth Circuit versus the Eleventh Circuit?

Somewhat. The Eleventh Circuit, in cases like Glasser v. Hilton Grand Vacations (2020), has been relatively receptive to human intervention defenses where agents manually dial each number. The Ninth Circuit historically read ATDS broadly, though post-Duguid those readings are limited. Neither circuit is a safe haven. Both examine actual per-call agent conduct rather than just system design.

Do ringless voicemail drops require human intervention to avoid ATDS classification?

Ringless voicemail drops delivered with no per-contact agent action have been found to lack sufficient human intervention in most courts that have examined them. The technology delivers a recorded message to voicemail automatically, with humans involved only at the campaign setup stage. The FCC has also indicated that ringless voicemails count as "calls" under the TCPA, so ATDS classification matters a lot for this channel.

If my vendor says their platform is not an ATDS, should I rely on that?

No. Vendor representations are marketing, not legal analysis, and they aren't binding on courts. Courts look at how a system actually functions, not what a vendor calls it. Request the system's technical documentation, understand what agents do at each step, and have counsel evaluate ATDS status independently. Vendor contracts that indemnify you against TCPA liability exist but give limited real-world protection against class certification and statutory damages.

How much are TCPA damages per call if I lose on the ATDS issue?

The TCPA provides statutory damages of $500 per violation, which courts have generally read as $500 per call or text. For willful or knowing violations, courts can triple that to $1,500 per call. In a class action covering thousands of calls, aggregate liability can reach tens of millions of dollars. That's why even small differences in dialer configuration generate heavy litigation.

Does having prior express written consent make the human intervention question irrelevant?

Mostly yes for autodialed calls to cell phones. If a consumer gives prior express written consent meeting the FCC's requirements under 47 C.F.R. § 64.1200, a call or text using an ATDS to their cell phone is permitted. Human intervention still matters if you lack consent, or if consent is disputed. Keeping clean, timestamped consent records is the most reliable way to avoid ATDS litigation entirely.

What FCC order should I read to understand the human intervention standard?

The FCC's 2003 Report and Order (CG Docket 02-278, 18 FCC Rcd 14014) is the primary source for the agency's human intervention guidance. It states that equipment requiring human intervention to operate each call is not an ATDS. The 2008 Declaratory Ruling in the same docket reaffirmed this. These predate Duguid but remain relevant for understanding how courts read the doctrine.

Can AI-assisted dialing tools claim human intervention even if AI chooses who to call?

This is open and genuinely unsettled as of mid-2026. If an AI system selects and queues contacts without human review of each one, most existing human intervention analysis would treat that as insufficient intervention, because the human didn't make the per-contact decision. If agents review AI recommendations and approve each one before dialing, the analysis is closer. Expect FCC guidance and new litigation on AI-assisted dialing in the next few years.

Sources

1. Cornell Law School, Legal Information Institute: 47 U.S.C. § 227: TCPA statutory text defining ATDS and setting damages at $500 per violation, tripled to $1,500 for willful violations
2. FCC, In the Matter of Rules and Regulations Implementing the TCPA, CG Docket 02-278, 2003 Report and Order, 18 FCC Rcd 14014: FCC 2003 ruling that equipment requiring human intervention to operate each call is not an ATDS; noted high abandoned call rates as a signature of predictive dialers
3. Supreme Court of the United States, Facebook, Inc. v. Duguid, 592 U.S. 395 (2021): Supreme Court held ATDS requires use of random or sequential number generator to store or produce numbers; systems dialing only from stored lists without such generation are not ATDSs
4. U.S. Court of Appeals, Eleventh Circuit, Glasser v. Hilton Grand Vacations Co., 948 F.3d 1301 (11th Cir. 2020): Eleventh Circuit held that dialing systems requiring human agents to manually dial each number, even with software assistance, did not qualify as ATDSs because of per-contact human control
5. U.S. Court of Appeals, Ninth Circuit, Marks v. Crunch San Diego, 904 F.3d 1041 (9th Cir. 2018): Pre-Duguid Ninth Circuit held that storing and dialing from a list could constitute ATDS; discussed human intervention as a factor in ATDS analysis
6. FCC Regulations, 47 C.F.R. § 64.1200, via the Electronic Code of Federal Regulations: FCC rule requiring prior express written consent for autodialed or prerecorded calls to wireless numbers for telemarketing purposes
7. FCC, In the Matter of Rules and Regulations Implementing the TCPA, CG Docket 02-278, 2008 Declaratory Ruling: FCC 2008 order reaffirming that predictive dialers constitute ATDSs; addressed automated message delivery without per-call agent action
8. FCC, Report and Order on Robocalls and Robotexts, CG Docket 21-402 (2023): FCC 2023 order tightening one-to-one consent requirements and identifying automated text campaigns as a priority enforcement area
9. Supreme Court of the United States, Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024): Overruled Chevron deference; courts now exercise independent judgment on statutory interpretation, giving FCC TCPA interpretations only Skidmore persuasive weight
10. FCC, Electronic Comment Filing System (ECFS), TCPA proceedings in CG Docket 02-278: FCC publishes rulemaking dockets for TCPA proceedings in the CG docket series, including human intervention and ATDS-related rulemakings
11. Federal Trade Commission, Telemarketing Sales Rule, 16 C.F.R. Part 310: FTC TSR governs telemarketing practices including use of predictive dialers and abandoned call rates, complementary to TCPA obligations