Outsourcing Insurance Follow-Up Calls: What Doesn't Transfer

You’re short-staffed, the leads keep landing, and someone on a sales call tells you the fix is easy: hire an outsourced call center, hand them the list, and the phones get covered. It sounds like the problem is solved. It isn’t, not the whole thing. Outsourcing your dialing can take the physical work off your desk. It does not, by itself, take the legal responsibility for that dialing off your agency, and that gap is where a lot of agencies get surprised later, usually by a demand letter instead of a phone bill.

This is about what actually transfers when you outsource lead follow-up calls, what doesn’t, what it costs using real published numbers instead of guesses, and how to vet a vendor, or an AI platform, so you’re not the one holding the bag for someone else’s dialing habits.

What “outsourcing your follow-up calls” actually means

The phrase covers three different arrangements, and agencies often don’t realize which one they’ve bought until something goes wrong.

A live-staffed answering service. A company like Smith.ai or a similar virtual receptionist provider answers your inbound calls under your business name, screens callers, and either handles simple requests or routes the call to your team. This is almost always inbound-only or lightly outbound (callbacks, appointment confirmations), billed per call or per minute with a monthly base.

A BPO seat lease. You rent one or more dedicated agents, typically overseas, who work your list on your schedule using your scripts, your CRM, and your phone numbers. This is the arrangement that looks most like hiring, because in practice it functions like hiring, just without the agency being the direct employer. It’s also the arrangement where the vicarious liability question is sharpest, because these agents are placing outbound calls, often at volume, on your behalf.

A staffing agency placing phone-only workers. Functionally similar to a BPO seat, but the workers may be domestic, and the staffing agency is the employer of record while you direct the day-to-day work. The legal analysis is closer to having an employee than a fully independent vendor, but the paperwork often doesn’t reflect that clearly.

None of these are wrong tools by default. The mistake is treating any of them as a complete answer to “I don’t have enough people to call my leads,” when what they actually solve is narrower than that, and what they don’t solve, consent, disclosure, documentation, stays exactly where it was.

Why agencies reach for outsourcing first

The reasoning is understandable. Licensed producer roles are slow to fill; a CSR-style phone role draws from a bigger labor pool but still takes weeks to hire and train. BLS’s May 2025 data puts the median wage for a customer service representative at $21.53 an hour, $44,770 a year, nationally. That’s before payroll taxes, benefits, a phone system, a manager, and the ramp-up time before a new hire is actually productive on your scripts and your carriers. Against that backdrop, a vendor who says “we already have trained agents, ready this week” sounds like it solves two problems at once: the empty seat and the training curve. If your agency’s staffing gap looks like the one covered in our piece on the 2026 insurance hiring shortage, outsourcing looks like the fast exit from that math.

It can be. The part that gets skipped is the paperwork that determines who’s actually responsible when the vendor’s agent, whether in Manila, Boise, or a staffing agency’s roster, makes a call that violates the TCPA, doesn’t honor an opt-out fast enough, or can’t produce a consent record when asked. That responsibility question doesn’t disappear because you didn’t personally dial the number.

The liability that doesn’t leave with the vendor

Here’s the mechanism, defined plainly, because the term gets thrown around loosely.

Vicarious liability means a principal, your agency, can be held legally responsible for the acts of its agent, the vendor, even though the principal didn’t personally commit the act. This is a general concept in the federal common law of agency, not something unique to insurance or to telemarketing.

The FCC applied this directly to the TCPA in a 2013 declaratory ruling, FCC 13-54, issued in the Dish Network matter (CG Docket No. 11-50, released May 9, 2013). The Commission held that a seller, meaning the business whose product or service is being sold, can be vicariously liable for a third-party telemarketer’s TCPA violations under federal common law agency principles, even where the seller itself never dialed a phone. The Commission was explicit that this isn’t unlimited: it rejected the idea that a seller is automatically liable just because a call happened to benefit it. The standard instead runs through classic agency-law tests:

• Actual authority. You told the vendor, explicitly or through your instructions, to make these calls this way.
• Apparent authority. You let the vendor represent itself to consumers as acting on your behalf, your business name, your script, your offer, even if you didn’t sign off on every detail of how it was executed.
• Ratification. You accepted the benefit of the vendor’s work after the fact, for example keeping and working a lead the vendor generated through a call that didn’t have valid consent, which can count as adopting the vendor’s conduct as your own.

Any one of those three can be enough to reach the seller, not just the dialer. In practice, that means the burden lands on your agency to show a court that the vendor was acting outside its authority, not on the vendor to prove it was acting within it. “We hired a call center and didn’t know what they were doing on our list” is a fact pattern that often points toward vicarious liability, not away from it, because it suggests you handed over authority without controlling how it was used.

This isn’t a reason to never use a vendor. It’s a reason to treat the vendor relationship the way you’d treat hiring: with a written agreement that spells out who does what, and documentation you can actually produce if a regulator or a plaintiff’s attorney asks for it.

What outsourcing actually costs in 2026

Real numbers instead of a range someone quoted you on a sales call. Smith.ai, a live-staffed virtual receptionist and answering service, publishes its pricing: the Starter plan runs $300 a month for 30 calls, the Basic plan $810 a month for 90 calls, and the Pro plan $2,100 a month for 300 calls, with per-call overage running from $11.50 down to $8.50 as you move up tiers. Extra features, appointment booking, CRM integration, bilingual coverage, run additional per-call or per-month fees on top of that base.

That’s for answering and routing. It’s not outbound speed-to-lead dialing the moment a form posts, and it’s not warm-transferring a qualified prospect into a real appointment slot. A BPO seat lease built for outbound dialing at volume is priced separately, and the quotes vary widely enough by agency, region, and included management overhead that there isn’t a single public number worth citing here without misleading you. What is worth citing is the shape of the comparison: live-staffed, per-call pricing in the range Smith.ai publishes, against the in-house wage math from BLS above, against a per-minute platform rate.

Run your own numbers against your own average handle time and call volume before deciding anything. A shop doing 40 calls a month looks completely different on this math than one doing 4,000.

A worked example, using published rates

Take an agency handling 500 calls a month between inbound answering and outbound follow-up. Using Smith.ai’s published Pro-tier structure as the answering-service benchmark, 300 calls are covered by the $2,100 monthly base and the remaining 200 fall into overage at $8.50 each, for roughly $2,100 + $1,700 = $3,800 a month, and that figure only covers answering and routing, not outbound dialing the moment a lead posts. Staffing that volume in-house with one full-time CSR at the BLS median wage of $21.53 an hour runs about $3,730 a month in wages alone, before payroll taxes, benefits, a phone system, or management time, and it still leaves outbound speed-to-lead dialing as a separate problem to solve, because one person can’t dial instantly on every new lead while also answering the phone.

On TheAffordableAI’s Single Account plan, the same 500 calls at a four-minute average handle time work out to roughly 2,000 minutes, or $400 at the $0.20-a-minute rate, plus the $200 monthly subscription, landing around $600 to $625 for the month depending on exact call length, covering both inbound and outbound in one system. None of these numbers include your setup time, your management overhead, or your own agency’s specific call mix, and none of them are a promise about what your agency will actually spend. They’re each vendor’s own published rate, run against the same illustrative call volume, so you can see the shape of the comparison before you price your own numbers against it.

The licensing wrinkle nobody puts in the contract

There’s a second risk that outsourcing conversations rarely surface: what happens when a BPO agent, trying to be helpful or trying to close, drifts from “let me confirm your information and get you on the calendar” into actually answering a coverage question. The moment a caller is quoting, binding, or advising on specific coverage, that’s regulated producer activity in most states, and an unlicensed BPO agent doing it, even accidentally, even with good intentions, creates exposure that has nothing to do with the TCPA. It’s an unauthorized-practice-of-insurance problem layered on top of the vicarious liability one.

This is a scripting and monitoring problem as much as a legal one. A tightly scoped script, confirm details, check basic eligibility, book the calendar slot, hand off anything substantive, keeps a vendor’s agents inside the lane they’re actually allowed to operate in. A vague script, or a vendor incentivized on conversion rather than clean handoffs, pushes agents toward exactly the kind of improvisation that creates this exposure. Ask a vendor directly what their agents are trained to say when a caller asks a coverage question they’re not licensed to answer, and ask to see the actual script, not a paraphrase of it.

The vendor-vetting checklist that actually protects you

If you’re going to outsource any part of your calling, in-house hire, BPO, answering service, or AI platform, this is the diligence that matters. None of it costs anything beyond the time to ask.

1. Get consent capture in writing. How does the vendor document that a lead gave prior express consent before an autodialed or AI-voice call goes out? Ask for the actual field or timestamp they log, not a verbal assurance that “we handle that.”
2. Require access to raw call recordings and scripts, not just summaries. A vendor who will only show you dashboards and call counts, not the underlying audio and the actual script being read, is a vendor you can’t audit if a complaint comes in.
3. Put liability allocation in the contract, not just an assumption. Who is contractually responsible if the vendor’s agent violates the TCPA while working your list? Indemnification language should exist and should be specific, not a vague mutual-hold-harmless clause.
4. Confirm opt-out handling and timing. Ask how fast a “stop” or “remove me” request gets honored across the vendor’s entire system, not just on the one call where it was said.
5. Ask whether agents disclose who they’re calling for. A consumer receiving a call should be able to tell it’s connected to your agency, not a generic, unnamed call center, both as a trust matter and because misrepresenting the caller can itself violate the TSR’s prohibition on deceptive telemarketing acts.
6. For Medicare-related calls, confirm CMS marketing rule compliance specifically. A generic BPO built for retail or SaaS lead gen usually isn’t built for the TPMO disclaimer, recording retention, and marketing rules CMS applies to Medicare Advantage and Part D calls. Ask directly whether the vendor has ever handled Medicare-regulated calling, not just insurance calling broadly.
7. Decide who owns the data afterward. If you end the contract, do you get your call recordings, transcripts, and consent records, or do they stay with the vendor? Get that in writing before you sign, not after you need it.

A vendor who answers all seven clearly, in writing, might be a reasonable choice. A vendor who waves off two or three of these with “don’t worry about it, we’ve got you covered” is telling you something true: you’re the one who’ll be covering it, later, without documentation.

Where a managed AI caller fits, honestly

TheAffordableAI doesn’t remove the liability question above. Nothing legitimately can, short of never delegating any calling to anyone. What it changes is where the visibility sits. Every outbound dial, every inbound answer, every disposition, and every consent capture lands in your own HighLevel CRM in real time, not in a vendor’s internal system you have to request access to after something’s already gone wrong. The scripts and consent language are set with your agency directly, not handed to an offshore team you’ve never spoken with. Full detail on what’s included, number warmup, spam defense, multi-calendar routing, is on the features page. You can also build a version of this yourself with a dialer and a BPO seat; plenty of agencies do, and the checklist above is exactly what to run through either way. If you’d rather hear it than read another comparison table, there’s a demo call on the homepage.

When outsourcing, or AI, is the wrong call

Some of this work genuinely needs a human with judgment, and neither an overseas BPO seat nor an AI platform is the right fix for it. A client disputing a claim, a nuanced coverage comparison across three carriers, a renewal conversation with someone who’s been with your agency for a decade, those need a licensed person who knows the account. Volume tools, human or AI, are built for the repetitive first-touch layer: dialing the moment a lead comes in, qualifying basic details, booking a real calendar slot, and getting a ready prospect to a licensed human fast. Pushing complex, relationship-dependent conversations through any high-volume tool, live BPO agents included, tends to produce worse outcomes, not better ones, because the tool is optimized for throughput, not judgment.

If your actual bottleneck is that your producers already have qualified leads sitting in their calendar and no time to close them, more first-touch calling capacity makes that backlog worse. That’s a closing-capacity problem, and no amount of outsourced dialing fixes it.

There’s also a volume floor worth being honest about. If your agency handles 40 or 50 calls a month, the fixed monthly base on most answering services or BPO seat leases, and the setup overhead of a managed AI platform, may simply not be worth it yet compared to your existing staff absorbing the load directly. Every option in this article, human or AI, has a volume where it starts paying for itself and a volume below which it’s solving a problem you don’t have yet. Know which side of that line your agency is on before you sign anything.

The compliance layer, regardless of who’s dialing

Automated, autodialed, or AI-generated-voice calls to a wireless number require prior express consent under the Telephone Consumer Protection Act, 47 U.S.C. § 227, full stop, regardless of whether the call was placed by an employee, a BPO seat, an answering service, or a platform like ours. Disclosure has to be clear, and an opt-out has to be honored immediately once it’s requested, in whatever channel it comes through. Medicare-related calling carries CMS marketing rules on top of that, including required disclaimers, and those rules don’t relax because the call came from a vendor instead of your own staff.

The FTC’s Telemarketing Sales Rule adds its own layer on the deceptive-practices side: a caller can’t misrepresent who they’re calling on behalf of, and sellers who use third-party telemarketers still carry obligations around recordkeeping and truthful representations under the rule. Call recording adds a state-law layer on top of the federal one. Some states require every party’s consent before a call can be recorded at all, which matters directly if your BPO vendor records calls by default as part of its own quality process; we’ve covered the recording-consent side of this in detail in our piece on California’s call recording rules, and the same question, does the vendor’s recording practice match the law in the state your lead lives in, applies to any outsourced calling arrangement, not just one state.

None of this is a reason to avoid outsourcing calling work. It’s the reason the vendor checklist above exists, and the reason “we hired someone else to make the calls” has never been a complete answer to a compliance question, whether that someone else sits in your office, overseas, or inside a software platform.

Price out your options with the actual published numbers above, run the checklist against whoever you’re considering, human or AI, and make the call that fits your volume and your risk tolerance, not the one a sales pitch made sound simplest.