agenthelp

Trust & Security

Last updated: June 28, 2026

The short version

Three promises, plainly:

  • Your data is yours — we don't sell it, and you can delete it anytime.
  • We never train AI on your customers' conversations.
  • Every AI phone call opens by telling the caller it's an AI and may be recorded.

1. Your data, walled off from every other business

Every business's data is separated by the database itself — a control called row-level security — not just by app code. Think of it like an apartment building: everyone's in the same building, but each unit has its own lock and no one else has the key. Another business literally cannot pull your records; the database refuses the access on every single query, automatically. It's not a rule we hope people follow.

2. Stored locked

Your data is encrypted in transit (TLS) and at rest (AES-256) — unreadable even to someone who somehow got the raw storage files. Passwords are stored only as one-way hashes (we can't read them), calendar connections are encrypted, and payment card numbers never touch our servers — those go straight to Stripe.

3. We don't train AI on your conversations

Your assistant runs on Anthropic's commercial API, which is contractually prohibited from training its models on your data — unlike the free consumer versions of ChatGPT or Claude, which can. Your customers' conversations are used to answer them, and for nothing else.

4. Who else touches your data, and why

We use a small set of trusted providers to run the service — for the AI, phone calls, texts, email, payments, hosting, and error monitoring. Each sees only what it needs, each is under a data-protection agreement, and none are data brokers. The full list, with what each one sees, is on our subprocessors page.

5. AI phone calls — what callers hear

If you use the AI phone line, every call opens by saying it's an AI and that the call may be recorded, before anything else. That's automatic and can't be turned off, and it's designed to meet the strictest U.S. call-recording laws, including all-party-consent states like California.

6. Your controls

  • Delete any conversation from your dashboard.
  • Erase a specific customer's records across your whole account (a “forget me” request).
  • Close your account anytime — that deletes your data; we don't hold it hostage.
  • Old data is cleaned up automatically on a schedule, so nothing piles up forever.

7. If something ever goes wrong

No system is perfectly secure. If we ever discover a breach affecting your customers, we'll tell you promptly, explain what happened and what we're doing, and help you respond.

8. Where we stand on compliance

  • Encrypted at rest and in transit; continuous backups with point-in-time restore.
  • Built for U.S. businesses and their U.S. customers; our AI-disclosure meets state AI and call-recording laws.
  • We follow CCPA principles and offer a Data Processing Agreement (see your dashboard settings).
  • Not for protected health information — agenthelp is not HIPAA-eligible and must not be used for PHI.

9. Talk to a human

Security questions? Email security@getagenthelp.com. Our Privacy Policy and Terms have the full detail.