Why the Line Between 'Informational' and 'Telemarketing' Is Blurry — and Expensive to Cross

Most businesses discover their calls were "telemarketing" only after a class action arrives. The reason is simple: regulators care about what a call is for, not what technology made it. A friendly reminder can be regulated as a sales call if the script drifts even slightly.

The FCC's rules cover three triggers, according to the agency's own telemarketing guidance: telemarketing calls, calls made with an autodialer, and calls using an artificial or prerecorded voice. Cross any one of those lines and the TCPA applies — regardless of intent. And since the FCC's February 2024 Declaratory Ruling, AI-generated voices count as artificial voices, with, as one industry analysis puts it, no carve-out for tech that mimics a live agent.

The price of misclassification is steep. TCPA damages run $500 to $1,500 per call, with no cap, and recent class-action settlements have ranged from $4.75 million to $19 million. One compliance analysis notes that 50,000 non-compliant AI calls at those rates "could sink a whole program." A campaign that feels harmless can become an existential liability.

The trap is the mixed-purpose call. Under the standard framework, informational calls like appointment reminders require only prior express consent, while telemarketing calls require prior express written consent — a signed disclosure naming the business. The practical rule from industry guidance is blunt: when in doubt, treat the call as marketing and get written consent.

That's why a renewal reminder with an upsell pitch matters so much. A reminder about an upcoming renewal is informational. Add "and while I have you, we're running a special on our premium tier," and the call quietly becomes telemarketing — triggering the stricter consent standard, DNC registry checks, and higher damages exposure. The same logic applies to onboarding check-ins that pitch add-ons or survey calls that end with an offer.

This is why disciplined campaign classification matters. My AI Call Center classifies every campaign by purpose before launch, applies the stricter standard to mixed-purpose scripts, and reviews list source and consent records before any calls go out. The key practices that keep a campaign on the right side of the line:

• Classify each campaign as telemarketing or informational based on script content, and document the rationale before launch.
• Keep upsell language out of reminder and notification scripts — or obtain the higher consent standard if it stays in.
• Verify that consent is seller-specific and topically related to the call, since the FCC requires consent to be "logically and topically associated" with the interaction that prompted it.
• Honor opt-outs immediately and log them, since revocation must be honored within 10 business days under FCC rules.

The Fifth Circuit's recent Bradford decision may have relaxed the written-consent requirement in some states, but as Holland & Knight notes, it does not eliminate TCPA exposure — callers must still show clear, direct, and unequivocal consent. Conservative classification remains the safest posture, because the cost of guessing wrong is measured per call, with no ceiling on the total.

The FCC's Current Definition: What Makes a Call 'Telemarketing'

The difference between a legal call and a $1,500-per-call violation often comes down to one question: does the FCC consider your call "telemarketing"? The answer determines which consent standard applies — and regulators have been rewriting that answer at a rapid pace.

Under the FCC's framework, telemarketing and sales calls require prior express written consent — a signed disclosure that names the business, confirms consent, and states that consent is not a condition of purchase. Informational calls, by contrast, require only prior express consent, which is often satisfied when a customer provides their number for that specific purpose. Appointment reminders, fraud alerts, and delivery notices fall into this lighter category, according to telecom compliance guidance. The FCC's rules cover three trigger categories: telemarketing calls, autodialed calls, and calls using an artificial or prerecorded voice, as outlined on the agency's telemarketing overview.

The stakes of misclassification are severe: TCPA damages run $500 to $1,500 per call with no cap, and recent class-action settlements have ranged from $4.75 million to $19 million, per industry analysis. When in doubt, treat the call as marketing and get written consent.

The one-to-one consent rule, effective January 27, 2025, tightened what "written consent" means. Under the FCC's 2023 rule, passed on a 4-1 vote, consent must:

• Authorize no more than one identified seller
• Be "logically and topically associated" with the interaction that prompted it
• Identify the specific telephone number consented to

The FCC was blunt: a consumer giving consent on a car loan comparison site does not consent to calls about loan consolidation. Consent is not transferrable or subject to sale to another seller.

The definition is also expanding on two fronts. The FTC's March 2024 amendments extended parts of the Telemarketing Sales Rule to B2B calls, with most provisions effective May 16, 2024, per a mid-year legal review. At the state level, Maryland now requires written consent for automated-dialer solicitations (effective January 1, 2024), Maine mandates scrubbing against the FCC's reassigned number database (July 16, 2024), Georgia eliminated the "knowing" violation requirement and allows uncapped class actions (July 1, 2024), and Mississippi effectively bans telemarketing for Medicare supplement plans.

One complication deserves attention. The Fifth Circuit's ruling in Bradford v. Sovereign Pest Control rejected the written-versus-oral consent distinction, holding the statute provides "no textual basis for imposing different forms of consent based on call content." That ruling binds only Texas, Louisiana, and Mississippi — other circuits and state laws may still require written consent, which is why interstate callers should stick with the stricter standard. This is exactly why My AI Call Center classifies every campaign by call purpose and reviews consent records before any launch — a mixed-purpose call, like a renewal reminder with an upsell, gets treated under the stricter telemarketing rule.

AI Voices Changed Everything: Why Your 'Human-Sounding' Calls Are Legally Robocalls

If your outbound calls sound convincingly human, you might assume they're legally treated like human calls. The FCC disagrees — and the distinction now carries five-figure per-call consequences.

On February 8, 2024, the FCC issued a unanimous Declaratory Ruling classifying all AI-generated voices as "artificial or prerecorded voices" under the TCPA. That means every outbound call using AI voice technology — real-time conversational agents, voice cloning, or LLM-driven bots — triggers robocall consent requirements, regardless of how natural the voice sounds.

The FCC was blunt about the practical effect: the law has no carve-out for tech that mimics a live agent. A call placed by a fluent AI agent with zero consent records is treated exactly the same as a robocall blasting a prerecorded pitch to strangers, according to compliance analysis of the ruling.

The stakes are not theoretical. TCPA damages run $500 to $1,500 per call with no cap, and class-action settlements between 2025 and 2026 have ranged from $4.75 million to $19 million. At that rate, a hypothetical 50,000-call non-compliant campaign could sink an entire program.

Disclosure rules are catching up fast. Texas now requires callers to disclose that a voice is AI within the first 30 seconds of a call, and a federal rule is widely expected to follow. The EU's AI Act goes further, requiring interactive AI systems to inform people they're talking to AI unless it's obvious from context.

Opt-out handling has hardened too. The FCC's February 2024 order established "per se" reasonable revocation methods — words like STOP, REVOKE, CANCEL, and UNSUBSCRIBE — that businesses must honor:

• Opt-outs must be honored within 10 business days, across all channels
• Callers cannot prescribe their own preferred revocation method
• A recipient can revoke consent mid-sentence, and the caller must stop before the next pitch begins

This is why list discipline matters more than voice quality. As industry guidance puts it, the decisive capability for an outbound AI program is not sounding human — it's respecting the human who answers. Defensible programs call only people with documented consent, verified before the first dial.

That's the same standard My AI Call Center applies to every campaign: list source and consent records are reviewed before launch, AI disclosure runs on every call, and keyword opt-outs like STOP and REVOKE are logged and honored immediately. Managed outbound calling campaigns for approved, permissioned lists start at 9¢ per connected minute — plan your campaign with a full quote before anything launches.

How to Classify and Document Every Campaign Before the First Call

The most expensive telemarketing mistake happens before a single number is dialed. TCPA damages run $500 to $1,500 per call with no cap, so a misclassified campaign can compound into serious exposure fast.

Classification starts with call purpose, not technology. Telemarketing and sales calls require Prior Express Written Consent — a signed disclosure naming the business and confirming consent is not a condition of purchase — while informational calls like appointment reminders generally need only Prior Express Consent, according to industry compliance guidance. Wait — that distinction has limits. In Bradford v. Sovereign Pest Control (February 2026), the Fifth Circuit held that the TCPA's text "provides no textual basis for imposing different forms of consent based on call content" — though that ruling applies only in Texas, Louisiana, and Mississippi.

The practical takeaway: classify every campaign during planning, write down the rationale, and apply the stricter standard to mixed-purpose calls. A renewal reminder that drifts into an upsell is a telemarketing call. The FCC's one-to-one consent framework requires consent to be "logically and topically associated" with the interaction that prompted it — the FCC's stated example: someone consenting on a car loan comparison site does not consent to robocalls about loan consolidation.

Before launch, verify each consent record against three checks:

• The seller is named — consent authorizes one identified seller, not a rotating cast of businesses.
• The topic matches — the campaign purpose relates to the interaction where consent was given.
• The phone number is identified — the record specifies where calls should be delivered.

Bought lists deserve special scrutiny. Consumer consent is "not transferrable or subject to sale to another seller," and a genuinely cold call — an artificial voice dialing a stranger with no consent record — can violate robocalling rules outright. Outbound calling places the burden of documenting the relationship on the company making the call, so a list without clear permission records is a liability, not an asset. This is why list and consent review sits at the center of how My AI Call Center runs campaigns: list source, consent records, and calling windows are checked before launch, and bought lists without clear permission records are flagged — and in most cases declined — before you spend anything.

When the classification is genuinely unclear, the industry rule of thumb holds: when in doubt, treat the call as marketing and get written consent. Written records also matter because the FCC classifies AI-generated voices as "artificial or prerecorded voices" under the TCPA, meaning AI-driven campaigns trigger robocall consent requirements regardless of how human the voice sounds. Even under the Fifth Circuit's more permissive reading, companies must still demonstrate "clear, direct and unequivocal consent." Documenting the basis for every call — before the first one — is what turns a compliance risk into a defensible program.

The Compliance Controls That Keep Outbound Calling Defensible

Knowing what counts as a telemarketing call only matters if your calling operation can prove it followed the rules. Outbound calling shifts the burden of proof to the caller — the company making the call has to document the relationship, not the person answering it.

That is why defensible programs run on a fixed set of baseline controls rather than good intentions. Industry guidance recommends six: consent verification, upfront AI disclosure, real-time opt-out suppression, calling-window enforcement, per-call recordkeeping, and escalation to a human agent. Each one exists because a specific regulation demands it.

The stakes are concrete. TCPA damages run $500 to $1,500 per call with no cap, and industry analysis warns that 50,000 non-compliant AI calls could sink an entire program. Class-action settlements between 2024 and 2026 have ranged from $4.75M to $19M. Compliance is not a legal abstraction; it is arithmetic.

Three controls deserve special attention:

• Consent verification — consent must name the specific seller and match the campaign's purpose. The FCC's one-to-one consent framework makes clear that consent is "not transferrable or subject to sale to another seller."
• Opt-out suppression — the FCC established "per se" revocation methods like "stop" and "revoke," and businesses must honor opt-outs within 10 business days.
• Per-call records — every call should produce a record of the calling basis, disclosure version, transcript, and suppression confirmation. The FTC's updated Telemarketing Sales Rule expanded recordkeeping to include call detail records.

Two operational rules turn these controls from paperwork into protection. First, retain records for the longest applicable period — industry guidance suggests at least 4 years, with 7 years the safer standard. Second, treat failed suppression write-backs as control incidents that block further dialing. If an opt-out does not write back to the list, the dialer stops — no exceptions.

This is where a managed, list-disciplined approach changes the game. My AI Call Center reviews list source and consent records before any campaign launches, flags bought lists without clear permission records, and runs only approved, permissioned, or reviewed lists. Nothing launches until the client approves the script, disclosure, and escalation path.

The result is that compliance stops being a legal gamble and becomes an operational checklist. Every call has a documented basis. Every opt-out is logged and honored immediately. Every campaign produces dispositioned contact lists, opt-out and DNC logs, and per-call records. As compliance practitioners put it, the decisive capability for an outbound program is not sounding human — it is respecting the human who answers.

Plan your next campaign with a free campaign review — managed outbound calling against approved, permissioned lists, from 9¢ per connected minute, with the full number known before launch.