FAQ

Quick answers to common questions. For deeper context, follow the links into the full docs.

General

Q: What's a Finn? A: An AI voice agent. It picks up the phone (inbound) or makes outbound calls, talks to the caller in natural language, and takes structured actions like booking appointments, qualifying leads, or transferring to a human. See Introduction.

Q: How is this different from a chatbot? A: Finns talk on the phone — real voice, real interruptions, real ambient noise. Chatbots type. The conversation model is similar; the surrounding telephony, latency budget, and voice quality matter much more.

Q: How fast can I get up and running? A: A first test call in about 15 minutes. A polished production campaign in a day or two. See Getting Started.

Q: What languages are supported? A: 30+ at varying quality tiers. Tier-1 (best): English (US/UK/AU/IN), Spanish, French, German, Portuguese, Mandarin, Japanese, Hindi. Tier-2: 20+ others with lower-quality voices and weaker STT. Check the voice library for your specific language.

Voice & quality

Q: How human do the Finns actually sound? A: For tier-1 voices, most callers don't realize they're talking to AI within the first 30 seconds. Some never realize. Some catch on by minute 2. Disclosure is recommended (and legally required in some jurisdictions) regardless.

Q: Can I clone my own voice? A: Available on Scale + Enterprise plans. You provide 5 minutes of clean recorded audio and a written consent statement; the platform creates a custom voice usable across your Finns.

Q: What's the latency? A: Average end-of-user-speech to start-of-Finn-response is ~400-700ms with default voices. Premium voices add ~100ms. Below 800ms feels natural; above 1.5s feels broken.

Reliability

Q: What happens if the LLM provider goes down? A: The platform fails over to a secondary model. Quality may degrade slightly during failover; calls don't drop.

Q: What's your uptime guarantee? A: 99.9% on Growth+, 99.95% on Enterprise (with SLA). Status: status.hirefinn.ai.

Q: What happens to in-flight calls if I pause or stop a deployment? A: They finish naturally. Pause/Stop only stops new calls from being dialed.

Compliance & privacy

Q: Is FinnAI HIPAA compliant? A: Yes, on Enterprise with signed BAA. Other plans are not BAA-eligible.

Q: GDPR? A: Yes — right-to-delete, data export, EU data residency available. See Integrations → Compliance.

Q: Are calls recorded? Can I disable recording? A: By default yes, for analytics and quality assurance. You can disable per-Finn under Call Settings, or globally under Settings → Compliance. Disabling recording also disables transcripts and post-call analysis.

Q: How long are recordings kept? A: 90 days default. Configurable up to 7 years for regulated industries.

Q: What about caller consent? A: Two-party consent is required in many US states (CA, FL, IL, MA, etc) and several countries. Add a recording disclosure to your Finn's welcome message; the platform provides templates.

Billing

Q: How do you bill? A: Subscription + per-minute AI usage + per-minute telephony. See Billing.

Q: Do I pay for failed calls? A: Telephony attempts are billed by the carrier and passed through. AI minutes are not consumed if the call didn't connect. So a failed call costs cents, not dollars.

Q: Can I cap my spend? A: Yes — per-deployment, daily, and monthly caps. See Billing → Budget caps.

Technical

Q: Can I use my own phone number? A: Yes — port it in, or SIP-trunk it. See Phone Numbers → BYO.

Q: Can the Finn look up data from my systems mid-call? A: Yes — via Webhook tasks in a Workflow. The Finn calls your endpoint, you return JSON, the Finn uses it in the response. See Integrations → Webhooks (Inbound).

Q: Is there a REST API? A: Yes. Auth, deploy, fetch results, manage audiences — full coverage. See Integrations → REST API.

Q: Webhooks for real-time events? A: Yes, with HMAC signing. See Integrations → Webhooks.

Q: Can I version-control my Finn configs? A: Yes — every save creates a version. Roll back from the Edit screen. Workflows can also be exported as JSON for git-tracking.

Performance

Q: Why is my Finn giving wrong answers? A: Usually one of: knowledge base too thin, conflicting source material, identity/guardrails not specific enough, or retrieval returning off-topic snippets. Debug via the Knowledge tab in the Data Extractor Sidebar to see what was retrieved per call. See Knowledge Base → Tuning retrieval.

Q: Why is completion rate low? A: Almost always a list-quality issue, not a Finn issue. Stale numbers, wrong country codes, calling at bad hours. See Audience → CSV gotchas.

Q: Why is my Finn hanging up too early? A: Usually max_idle_duration is too short, or the welcome message is set to defined but empty. See Creating a Finn → Call settings.

Q: Calls cut off at 5 minutes — bug? A: Default limit_call_duration is 5 minutes. Raise it if your use case needs longer. Be aware: longer caps = higher per-call cost ceiling.

Account

Q: Can I have multiple users on one account? A: Yes, on Growth+. Invite via Settings → Team. Per-user roles: Admin, Operator, Viewer.

Q: SSO? A: SAML 2.0 + SCIM on Enterprise. See Integrations → SSO.

Q: How do I cancel? A: Settings → Plan and billing → Cancel. Effective at the end of the billing period. See Billing → Cancelling.

Still stuck?

• Read the full docs.
• Email support@hirefinn.ai.
• Talk to a Finn (the irony) — the chat bubble bottom-right is itself a Finn answering support questions.
• For Enterprise: dedicated Slack channel with your CSM.