# Answer the lead after hours. Call aged leads. Book the LO’s calendar. Take the payment at 9pm.

*Voice Agents*

> Origination agents call web leads back in seconds and book loan officers’ calendars. Servicing agents verify identity, take payments, change due dates, quote payoffs and handle hardship intake — then warm-transfer if needed. Every conversation is scored afterwards against your rulebook.

## The desks Voice Agents powers

- **[Loan Officer Assistant](/for-lenders/loan-officer-assistant)** — Answers every lead in seconds and books the LO
- **[POS](/for-lenders/pos)** — The borrower’s front door, inside the system
- **[Document Concierge](/for-lenders/document-concierge)** — Gets the full document set before underwriting asks
- **[Escrow Manager](/for-servicers/escrow-manager)** — Analysis, shortages, and the explanation that is borrower-friendly
- **[Collections Agent](/for-servicers/collections-agent)** — Early-stage delinquency, compliant with Reg F and TCPA
- **[Loss Mitigation Agent](/for-servicers/loss-mitigation-agent)** — Complete packages, evaluated inside the Reg X clock
- **[Payoffs Manager](/for-servicers/payoffs-manager)** — Quotes, good-through dates and the release clock

## From the first ring to the action taken

1. **Connect** — Sei sits on top of the telephony and CRM you already run — Genesys, RingCentral, Twilio, Velocify, Salesforce — and reads the loan file from your LOS.
2. **Converse** — The agent handles the call end to end against your scripts and the rules that apply to it, checking frequency and disclosure requirements as it goes.
3. **Act, then hand off** — Actions execute downstream — payment taken, due date changed, appointment booked, ticket raised. Anything outside the agent’s rules, result in warm-transfers to your team with a summary.

## Proof

A top-25 independent mortgage bank put Sei on inbound and outbound origination calls.

- **9.2×** — more appointments set on average
- **+12%** — lift in lead conversion
- **24/7** — inbound coverage

Appointments set with loan officers rose **9.2× on average** — enough that several said they now needed to block hours out of their own calendars.

[Read the case study](/case-study/sei-x-ca-mortgage-lender)

## Demo

[Voice Agents on a borrower call](https://www.youtube.com/watch?v=tJ73Ztrmtw4) — A full conversation, start to finish — verification, the request, the action taken, and the hand-off.

## Conversational multi-channel AI

### Origination calls

- Inbound purchase and refinance enquiries, 24/7
- Hand-off into the point-of-sale your borrowers apply in
- Speed-to-lead callbacks on web and portal leads
- Pre-qualification questions and criteria screening
- Loan-officer appointment booking
- Application status for a borrower already in the pipeline
- Chasing the document the file is waiting on

### Servicing and collections calls

- Identity verification to your standard
- Payments and payment arrangements
- Due-date changes
- Payoff and reinstatement quotes
- Escrow and PMI questions
- Hardship and loss-mitigation intake

### Rules the agent works inside

- FDCPA, TCPA and UDAAP
- Regulation F — the 7-in-7 call-frequency limit and limited-content messages
- ECOA and Fair Housing
- Your call scripts, disclosures and escalation policy

## What sets it apart

- Built against FDCPA, TCPA, UDAAP and CFPB enforcement actions
- Reg F-aware outreach, including the 7-in-7 frequency limit and limited-content messages
- Runs across the loan lifecycle: the lead at the front, document follow-ups and the payoff at the back
- Every call scored afterwards by Sei’s own monitoring — against TILA, RESPA, TRID and UDAAP as well as your scripts

**9.2×** — more appointments booked

## When it is unsure

If a caller asks something outside the agent’s remit, or identity cannot be verified to your standard, the agent warm-transfers to your team with a summary.

## FAQ

**What happens when the caller asks something the agent cannot handle?**

The agent warm-transfers to your team with a summary of the call so far rather than guessing — and it does the same if identity cannot be verified to your standard. It does not improvise an answer it is not confident in.

**Do we have to replace our dialer or contact center?**

No. Sei runs on top of the telephony and CRM you already have — Genesys, RingCentral, Twilio, Salesforce and HubSpot among them — and reads and writes the loan file in your LOS. Custom integrations are handled during onboarding by a dedicated account team.

**Which rules are the agents built against?**

FDCPA, TCPA and UDAAP, plus Regulation F for collections outreach — including the 7-in-7 call-frequency limit and the limited-content message — alongside ECOA, Fair Housing and your own scripts and disclosure requirements. The rule set is yours, and the agent reasons against it rather than against a general model’s memory of lending.

**Who checks the calls the agent handled?**

Every one of them is scored afterwards by Sei’s Call Monitoring — against TILA, RESPA, TRID and UDAAP alongside your own scripts and SOPs — with the violating moment cited to its timestamp. All of them, not a sample, and you are not taking the agent’s word for its own compliance.

## The regulatory ground under borrower calls

- [Regulation F for AI Voice Debt Collection: The 7-in-7 Ledger, the Limited-Content Message, and the Per-Debt Architecture](/blog/regulation-f-ai-voice-debt-collection-7-in-7-limited-content) — The CFPB's Regulation F call-frequency presumption, post-conversation wait, limited-content message, and channel opt-out rules are mechanical and per-debt. The agent design we run so an AI voice program does not inherit a 1006.14 harassment finding from a dialer it replaced.
- [TCPA Compliance for AI Voice Agents in Mortgage and Bank Outreach: A 2026 Field Guide](/blog/tcpa-ai-voice-mortgage-banking-2026-field-guide) — What the FCC's 2024 declaratory ruling, the one-to-one consent rule, and the Mortgage One class action mean for AI voice outreach — with the consent stack, suppression rules, and audit pack regulated lenders need.
- [The SAFE Act Line for AI Mortgage Assistants: When Quoting a Rate Becomes Loan Origination and What NMLS Cares About](/blog/safe-act-ai-mortgage-assistants-nmls-licensing-line) — The SAFE Act and Reg G/H were written for human MLOs and the definition of 'loan originator' bites on any person who takes an application or offers or negotiates loan terms. AI assistants on the mortgage intake desk straddle that line without realizing it, and the state regulators that examine NMLS are starting to ask. The boundaries we hold the agent to so the institution does not have to sponsor a license for software.

---

_Source: [https://www.seiright.com/platform/voice-agents](https://www.seiright.com/platform/voice-agents) · Sei AI_