What Borrowers Think When an AI Calls Them
Will an AI call annoy your borrowers or hurt their trust? What earns trust on the call, what breaks it, and how to hear the call yourself before you decide.
Most borrowers judge an AI call the way they judge any call: who is it from, is it worth my time, and will they leave me alone if I say no. Trust holds when the call names your company right away, doesn't pretend to be a person, comes at a reasonable hour, makes one clear offer, accepts 'no' the first time, and hands interested borrowers to a real person at their own branch.
Most borrowers judge an AI call the same way they judge any call. Who is this? Is it worth my time? Will they leave me alone if I say no?
Get those three right and the call feels like hearing from a place they already know. Get them wrong and it feels like spam, whether a machine or a person placed it.
So the real question isn't whether borrowers like AI. It's whether your call earns the next 30 seconds. That part is in your control.
The fear is reasonable
If you run operations at a lender with branches, you've probably had this thought: our customers know us. They know the manager by name. What happens when a robot calls them?
That's not a silly worry. It's the right one.
Your branches run on relationships. A borrower who paid off a loan two years ago might still wave at the manager in the grocery store. You don't want to trade that for a few extra calls.
And the public mood isn't warm. In a June 2025 survey of 5,023 U.S. adults, Pew Research Center found that 50% of Americans are more concerned than excited about the growing use of AI in daily life, up from 37% in 2021. About six in ten said they want more control over how AI is used in their lives.
Two honest caveats. Pew asked about AI in general, not phone calls. And the same survey found 76% say it's extremely or very important to be able to tell whether pictures, videos and text were made by AI or by people. That was about content, not calls either. But the direction is clear enough: people don't like being fooled, and they want a say.
That tells you how to design the call. It doesn't tell you not to make it.
What a bad call sounds like
Picture a borrower who paid off an installment loan in the spring. It's a Tuesday, 11:10 a.m. She's on her break.
Her phone rings from a number in another state. She lets it go. It calls again twenty minutes later. She picks up.
A friendly voice asks for her by first name and wants to know how her day is going. No company name. It chats. It sounds like a person, almost. Then it mentions a loan.
She asks "who is this?" It answers with a script. She says "not interested." It asks if she's sure.
That call didn't fail because it was AI. It failed because it hid who was calling, pretended to be something it wasn't, and didn't take no for an answer. A person reading the same script from a call center would have done the same damage.
Now the good version.
Six things that keep a borrower's trust
1. Say who it's from, right away. The first sentence names your company and the branch she knows. This is also the rule: the FCC's 2024 ruling, FCC 24-17, confirmed that AI-generated voices count as artificial voices, and artificial voice messages must clearly state the identity of the business responsible for the call at the beginning.
2. Don't pretend to be a person. A good agent sounds natural. It doesn't claim to be human. If she asks "am I talking to a real person?", the answer is a plain no, followed by an offer to connect her with someone at her branch. Getting caught pretending costs far more trust than being honest up front.
3. Call from a number she can place, at an hour that makes sense. In a July 2020 survey, Pew found 80% of Americans generally don't answer calls from unknown numbers. A local number with honest caller ID helps. Timing matters too. Federal rules put telemarketing calls between 8 a.m. and 9 p.m. her local time, per 47 CFR 64.1200. We'd go tighter: only call while her branch is open.
4. Make one clear offer. Not a pitch. One sentence she can say yes or no to. "You paid off your loan with us, and you may qualify for a new one. Would you like to talk to the branch about it?" Anything about rates, amounts or approval belongs to a person, not the agent. We cover that line in what an AI voice agent should do at a lender.
5. Take "no" and "don't call me" the first time. No second ask. "Don't call me" goes straight onto the same internal Do-Not-Call list your branches use, so a loan officer doesn't call her next week. The same federal rule says a request to stop calls must be honored within a reasonable time, no more than ten business days. We'd aim for the same day.
6. Hand her to a real person at her own branch. If she says yes, the agent transfers her live to her local branch while it's open. Not a callback tomorrow. Not a call center in another state. The branch she walks into. That's the moment the AI gets out of the way, and it's the part borrowers remember. Here's why a live transfer beats a callback.
This is educational, not legal advice. Review your scripts, consent and disclosures with your own compliance counsel. Our compliance checklist lists what to bring.
When a person should make the call instead
Being honest about this matters more than selling you on AI.
If your list is small, have your branch staff call it. A manager who knows the borrower by name will beat any agent.
If the conversation is sensitive, like a hardship, a dispute or a complaint, keep AI off it. Take those people off the list before the first call.
If a borrower has already told you they don't want calls, they stay off. Every time.
AI calling fits the volume work: thousands of past customers nobody has time to reach, with a simple first question. It works because the relationship part still happens at the branch, with a person. If you're comparing vendors, these 12 questions include asking what their agent will never say.
What we've seen with a real lender
Sunset Finance is a network of consumer lending branches in Georgia. They had 43,000 past customers who had already borrowed once, sitting unused. The offer was a pre-approved loan of up to $1,000.
Their voice agent, Tommy, placed 13,159 calls in the first 14 days of the pilot. About 100 of those borrowers chose to be transferred live to their own local branch in those two weeks, across 10 branches. It's 224 live transfers and counting since, and the program has grown from 10 branches to 28.
We won't tell you every borrower loved the call. We don't have a survey that says so. What we can show you is what borrowers did: when the call was clear about who it was from and what it offered, some of them said yes and asked to talk to their branch, right then.
Hear the call before you decide
You're probably still picturing the bad version. That's fair. The only way to settle it is to hear the good one.
Meanwhile, the list keeps aging. The borrower who paid off in the spring hears nothing from you this month, and next month, and the relationship you're protecting gets quieter on its own. Silence protects nothing.
If you want to hear what your borrowers would hear, the AI will call you. Leave your number.
Questions lenders ask
Will an AI call annoy our past customers?
It can, if it hides who's calling, calls at bad hours, or won't take no for an answer. A call that names your company, makes one clear offer and stops when asked is much closer to a normal branch call.
Should the AI say it's an AI?
It shouldn't pretend to be a person. Pew Research Center found half of Americans are more concerned than excited about AI in daily life, so getting caught pretending costs more trust than being plain about it.
Does the call have to say who it's from?
Under the FCC's 2024 ruling on AI voices, artificial or prerecorded voice messages must clearly identify the business responsible for the call at the beginning. Review your own calls with counsel.
What happens when a borrower is interested?
The best version transfers them live to a real person at their own local branch while it's open, instead of promising a callback.