The FCC just published official guidance confirming that TCPA (Telephone Consumer Protection Act) applies to AI-generated voices. Statutory damages remain $500–$1,500 per call. Read the full FCC document here.
This changes everything for AI voice agencies. Up until now, compliance was a grey area. Agencies built on APIs, routed calls through whoever had the cheapest infrastructure, and hoped regulators wouldn't come knocking. Today, regulators came knocking. And the cost of non-compliance just got real.
Why This Matters for AI Voice Agencies
If you're running an AI voice agency, you're now liable for every call your agents place. Not just the software provider. You.
Here's the liability chain: You use an API provider (Retell, VAPI, Twilio). That provider isn't the one calling—your client is, on your recommendation, using your agent. If the call isn't compliant, the FCC comes after your client first (they own the business relationship). Your client then comes after you (you sold them the tool). You then try to come after the API provider (they usually have TOSs that shift liability back to you).
The escape hatch: You need infrastructure that handles compliance automatically. A2P compliance gating, consent tracking, do-not-call registry checks, state-level consent rules (California, New York, and 14 other states have heightened standards). If you're duct-taping together Retell + GoHighLevel + Zapier + a custom database, you're not tracking any of this. You're exposed.
The good news: compliance isn't optional anymore, which means it's now a market category. Agencies that move first get to own the compliance narrative with their end-clients. Agencies that wait get undercut by someone who says, "Our platform handles A2P compliance for you. $30 per submission. Your risk, our infrastructure."
What We're Doing at Hermes About It
Hermes was built with compliance from day one. Every agent deployed on Hermes runs through A2P compliance checks before it can place a call. We manage the submissions to carriers (AT&T, Verizon, T-Mobile). You don't touch the paperwork. Your clients don't get sued.
Here's what's included:
- Do-not-call registry checks (DNCL) per call
- A2P carrier pre-approval + submission handling ($30/submission, pass-through cost)
- State-level consent tracking (California opt-in, New York DNC, etc.)
- Consent recordings + audit trail (admissible in court)
- TCPA liability insurance partnerships (coming Q4 2026)
You deploy an agent on Hermes for $149/month (Starter). That includes 300 minutes of outbound calling, compliance included. You don't hire a lawyer. You don't debug A2P submissions. You hand off the regulatory risk to us.
Your end-clients see this as competitive advantage. "We use Hermes" becomes a selling point: "Our voice agents are FCC-compliant. We handle the risk. You handle the ROI."
Action Steps for Agencies Affected
If you're currently running AI voice agents on Retell, VAPI, or any API provider, here's what to do this week:
- Audit your current outbound calls. Log into Retell/VAPI dashboards. How many calls did you place in the last 30 days? Export the list. Check: Are any of these going to numbers on a do-not-call list? Are you recording consent? If you don't have answers, you have a problem.
- Document your consent flow. For every call, you need to prove the contact agreed to be called by an AI. If you don't have this, go back to your end-clients and ask: "How were these leads sourced? Did they opt in?" If the answer is "they're from a rental list" or "cold outbound," you're non-compliant. Fix the sourcing rules in your agent logic.
- Check your state-specific rules. California requires prior express written consent. New York requires a no-call list check on every call. Texas, Florida, Pennsylvania have their own rules. If you're deploying nationwide, you need state-level logic in your agent. Most API providers don't have this built in. Hermes does.
- Plan your compliance infrastructure. You have three options: (a) Build it yourself ($50K+ in dev time, hiring a compliance consultant), (b) Use a compliance-first platform like Hermes ($149/month, handles it for you), or (c) Keep using APIs and hope you don't get audited. Option (c) is what you're doing now. The FCC just made that a bad bet.
- Talk to your liability insurer. Tell them you're running AI voice agents. Ask: "Are we covered if we get an FCC complaint on outbound calling?" If they say no, or if they want to charge a rider, that's your signal to move to a platform with built-in compliance. It'll be cheaper than the insurance premium delta.
The Math on Compliance vs. Non-Compliance
Let's say you're an agency with 10 clients. Each client runs 1,000 calls/month. That's 10,000 calls/month, 120,000 calls/year.
If even 1% of those calls are non-compliant (1,200 calls/year), and the FCC audits, that's 1,200 calls × $500/call minimum = $600,000 in liability. That's your company. And that's before your clients sue you for putting them at risk.
A2P compliance on Hermes is $30/submission. You submit once per agent. If you have 10 agents, that's $300/year. Your Hermes hosting is $149/month × 12 = $1,788/year. Total: $2,088/year for full compliance cover.
The bet is whether $2,088 in annual compliance cost is less than the expected value of an FCC audit. (Spoiler: it is by orders of magnitude.)
Competitive Signal: What This Means
Synthflow and other builders are going upmarket (enterprise, $3K+/month). Retell and VAPI stay API-only (they shift liability to you). Hermes is building agency infrastructure that assumes compliance from day one.
This FCC guidance just handed agencies a business model: become the compliance layer. Charge your end-clients a markup for "compliant deployment." On Hermes, your cost is $149/month. You can charge $299/month and pocket $150/month per client (80% margin). That's the move.
FAQs
Does the FCC guidance apply to inbound calls too, or just outbound?
The guidance primarily covers outbound calling (initiating calls). Inbound is simpler: when a contact calls you, you can use AI from the start of the call. No prior consent needed. Outbound is the blocker. That's where compliance infrastructure matters.
If I use Hermes, am I automatically FCC-compliant?
Hermes handles the technical compliance (DNC checks, carrier pre-approval, state rules). But you still own the sourcing. If you buy a cold list and run agents against it, that's not compliant, and no platform can fix that. You need contacts that have opted in. Hermes makes sure your opted-in contacts reach them legally.
What if my end-client insists on using Retell or VAPI?
That's their choice. But tell them: you're liable for compliance. If they want you to be liable, you need to charge a compliance fee ($50–100/month). Or you can recommend Hermes as a white-label platform where they own the compliance risk. Either way, the cost of compliance is now visible. Don't absorb it into your service fee.
Next Steps
Start here: Join the Hermes beta and we'll walk you through the compliance setup for your first agent. Everything is built in. No custom development, no lawyers, no excuses.
If you want to compare Hermes to your current stack: See our competitor comparisons (Retell, VAPI, Synthflow). We'll show you exactly what you're missing on compliance.
Pricing starts at $149/month for Starter (300 minutes, 3 workspaces). A2P compliance is $30 per agent submission. Everything else is included.
The FCC just made compliance non-negotiable. The agencies that move first win. The rest get audited.