Skip to content
HomeServicesAboutIndustryPricingPodcastREDOJobsBook a Call
Advice from a Call Center Geek

Bots Calling Bots: Is Your Contact Center Ready for AI Customers?

Tom Laird · October 28, 2026 · 15 min

Listen
Apple Podcasts Spotify YouTube
Key Takeaways

What Tom covers.

Takeaway 1

Gartner predicted 20% of inbound contacts would come from machine customers by 2026; Tom estimates it is around 6% today

Takeaway 2

He expects more than half of call volume could be some kind of AI bot within 12–18 months

Takeaway 3

Good bots will identify themselves and dangerous ones won't, so the IVR has to become the router that unmasks them

Takeaway 4

Bots don't mind waiting, so Erlang C, abandon rate and service level may not work when bots sit in the queue

Takeaway 5

Liability and authorization are unresolved; some banks such as JP Morgan are refusing to deal with bots

Takeaways are paraphrased summaries of what Tom says in the episode, not direct quotes.

Show Notes

What happens when AI bots start calling your contact center on behalf of your customers? In this live solo episode, Tom shares his honest, unscripted thoughts on machine customers: bots asking for discounts, checking balances, negotiating bills and canceling subscriptions.

He thinks that in the next 12 to 18 months, more than half of your call volume could be some kind of AI bot, and he walks through the questions nobody has answered yet.

In this episode:

- Gartner's prediction that 20% of inbound contacts would come from machine customers by 2026, and where things actually stand

- Good bots introduce themselves and dangerous ones won't. How do you unmask them in your IVR?

- Routing: bots don't care if they wait, humans do. Should AI talk to AI while humans talk to humans?

- Why Erlang C calculators, abandon rate and service level break when bots sit in the queue

- The legal mess: who's liable when a customer's bot agrees to something the customer didn't want?

- Which industries get hit first: retail, internet and energy, and SaaS subscriptions

Tom doesn't claim to have the answers yet. If you hear people say they do, question it.

Host

Tom Laird

Tom Laird hosts Advice from a Call Center Geek and is the CEO of Expivia in Erie, PA. Tom also founded OttoQA (AI call QA).

Transcript
Read the full transcript

Machine-generated transcript, lightly cleaned. It may contain transcription errors; timestamps are approximate and untranscribed gaps are marked.

00:00Hi guys, Tom Laird here and you've heard me break down contact center technology on the podcast time and time again. Now it's time for me to help you with your challenges. As your contact center technology advisor through our new company, Expivia Digital, I can help you with things like election, any type of AI implementation, and nice studio and integration services. It's the same honest advice you hear on the geek applied specifically to your operation. Visit ExpiviaDigital.com to get started. That's ExpiviaDigital.com. Welcome back, everybody, to another episode of advice from a call center geek, the call center, contact center podcast, we try to give you some actionable items, take back in your contact center, improve the overall quality, improve the age and experience, hopefully improve your customer experience as well. How has everybody been doing? We are living in a crazy time right now. I know I see my eyes, I record this is kind of happening and going on in Orlando.

00:52I am not there. Uh, just, just too much going on here. And again, I think the amount of money that it costs for a booth, I'd rather kind of throw into some of our companies, just not finding the ROI on that anymore. But I want to talk about the big topic that is going on in the contact center space. I spoke a ton about this, and I think some people thought I was crazy, and I'll be honest, I got a lot of this at the nice world, there used to be nice interactions. This was one of the main topics, and I came back, did a giant post on this, and kind of got a little bit of poo-pooed on it. But now. Now, if you look on LinkedIn, this is all anybody is talking about. What is going to happen when AI and bots start to take over the customer experience, start to work on behalf of customers, your customers, the customer experience calls call the to get a see if they can get a discount. What is their account balance, transfer money, buy a product. It's it's happening right in front of our eyes right now. And I don't know how prepared all of us are for it. Again, most of the most of the things that you're going to see on LinkedIn are literally a talking that they just posted, ask ChatGPT for some information and they're throwing stuff at you. I want to throw my actual thoughts at you of what I that's going to happen, how to deal with it, and some things that we need to think of, because I don't have the answers to a lot of this stuff. It's also brand new, and that's what makes it fun. And it also creates opportunities, I think, for a lot of organizations to differentiate how you're going to handle this, because there's some organizations, and some, I think it was JP Morgan, and don't quote me on that. That said not, we're not, we're not going to deal with it. We're not accepting customer facing bots into our IVR, into our support model. And again, I think that that is maybe a wise decision to do right now, but it's not a long-term wise decision. So let's kind of get into this a little bit.

03:04Again, I'm live on LinkedIn, live on Facebook, on X. Please throw in any questions you have. I'd love to kind of make this a little bit more interactive. I'd love to hear some of your questions with this stuff. Everybody kind of starts with the Gardner quote. I'm going to do that. I'm not a massively huge Gardner fan. A lot of the things I think that they do is, I don't know, I disagree with a lot of stuff, but I think that this is starting to see kind of what you're starting to see. So they said that about 20% of inbound contacts will come from machine customers by 2026. Well, it's 2026. We're not there yet. But I think it's more like the stat that I'm seeing about 6% of AI is actually being used to contact companies. And it's probably maybe a a little bit under that with the amount of customers.

03:54But we're seeing bots, customer facing use bots coming out with Metamuse. I mean, GrokBot is absolutely amazing. I went from using Cloud 99 percent of the time to I'm almost fully on GrokBot, with a ton of different bots created, API connections and I don't know if you've seen what we've done with CXone, but you can just create an API connection with GrokBot into CXone to pull reports, make them look beautiful, they're there in 30 seconds, no more clicking on filters and doing all that jazz. Same thing is happening with a lot of the tools that we're utilizing. So this is starting to be programmed into human beings. I think that when you look at Google, and they're growing up, they're kind of that ask for me. So ask for me is kind of a test here in in the States where you can basically click the button, ask questions and it will basically make a phone call for you.

04:53What are your hours of operation? What is the cost of this? It will be working kind of on your behalf. So again, I think we've started to see even on the chat side and moving out of the voice side bots that are negotiating bills, they're canceling subscriptions. And how do we handle this? It's difficult. And again, like I said, I don't think anybody really has the answers and I read about this stuff all the time. And I think the more I read, the more I'm getting confused. So, let's talk about how many of these things are coming, right? And I don't think anybody really knows. And that's the problem. But I think in the next 12 to 18 months, I would say that over half of your call volume is going to be some type of AI bot contact you. depends what industry that you're in. Some will still remain very human, but if you're in retail, I mean, you better be prepared for this stuff. If you're in the FCC world of the internet, and energy, and those kinds of companies, you're going to get... Because people are going, I don't care, it's not going to really... No skin off my tail, and hey, you renegotiate, trying to renegotiate this. That's going to happen a ton. How are we going to really deal with that. And again, I think, again, we start to kind of get a little bit deeper.

06:15How do we know who a bot is, right? They're so difficult to find. And I think the quote unquote, good bots, they're going to introduce themselves. They're going to say who they are, but the dangerous ones won't. And we know about how much fraud is going on now. That's the scary part with kind of what is going on, I think, with some of this and why the JP Morgan's are kind of saying, Hey, we need to kind of put out put a hold on this. But there's lots of I think ways and CX one I know for one is trying to develop a way I'm trying to think through, how can we develop ways To not have, tell us who they are, but really unmask them when they come into RVR. And I kind of like what nice said to, you don't have to have voice facing AI, but you need to have kind of that, that porch, right where they're the first greeting, right is, is, is AI. And then that AI can can hand off to a human, maybe hand off to another bot that you have through an API or some type of other connectivity that will handle different types of interactions that we're dealing with now. How do we route these things? That's the question, question that I asked on LinkedIn and got a really good response. Think about this, bots don't have to, they don't care if they wait. Human beings do. If you don't have different ways for for your AI, for AI customers to go to talk to an AI bot that you have. And I don't even know if that's the right decision. If all you have is human, right, do you route them differently? If you can find out who they are, do you let the humans come first? I mean, I think we are at the world where AI is going to have to start to be talking to AI, because the volume will just be too big for human beings. But I think the human beings should be still talking to human beings. I think that's the divide we're seeing. I don't think we're going to see the, you know, that the AI takes over the human calls. I think, I think that's the new question, right that we're Thanks for watching!

08:42That that those interaction those experiences are happening I Think you know. You know the other thing that that we have to think about is what stats are important, right? You know the Erlangs and in kind of the Erlang see calculators. They assume people hang up They assume these these length of time incidents, but bots don't right a band and rate can drop, you know that by the queue explodes like Does a bot count toward a service level I I don't know. Are there two different metrics that we need to have for going to bot versus bot versus humans? Do we need to start to talk on those? I think the QA score still comes into play. But again, you know how I feel about the overall score. I don't think it's that important anymore. Call types are more important. The actual interactions are more important. I think you still need a QA or AI.

09:36There's a lot that I think we need to figure out the routing of it though I think is the first part and who actually He's going to handle the stuff. And then we get. It's a huge kind of legal issues. You know, who's liable when a customer's bought agrees to do something maybe they didn't want. And that's the confusion that's happens now. What if they, you know, made a mistake and instead of downselling, selling, they. They upsell then, then they're going to the human beings going to call back the company and be mad. that's gonna have to be worked out. Again, is there some type of authorization that has to happen that the customer will have to still be on the phone with them? Can they take that authorization back? The legal aspect of this is so overwhelming and confusing that I think, again, while I think the 12 to 18 month, 50% of all interactions will be kind of bot based, this is the one thing that can be kind of that, we're not thinking about this enough, and it's going to, it's never going to really catch on because of it. But I think there's still things that that's, it's not going to matter as much right when we talk about, you know, Reservation and and verification. I mean, I don't know what doesn't but you know There could be some type of calls with that but the process is not in place to really handle this Yeah, the the validation of in Marney I totally agree that's kind of the right the the validation for having a malicious entry for having someone who is Faking who is deep faking It's, it's, it's really, really hard to kind of start to figure out how that, that will come into play, and again, I know it keeps saying it, but like JP Morgan, they've said, absolutely not. we're not handling any of these, but there's other people were saying, I won't even deal with a company now that won't deal with my bot or my kind of my little helper, my AI helper.

11:46so it is, it is an interesting world. And again, we are trying to think this through very heavily here at Expivia, trying to think through the security aspect of it. We want to be able to offer service to these if a customer wants them to be. You know this is the thing that's keeping me up at night right now is really coming through. How do we validate? How do we authenticate? How do we Right? Don't quote me on that, but I think those three are with Muse right now trying to kind of come up with some protocols for this But it's it's a it is definitely an interesting kind of kind of world, you know, the other thing is If you put a bot on hold, you know If you could just tell it hey every time that you go on hold just call back go to the front of the queue try to get in front of the queue, but there's gonna be so much kind of manipulation that But it will be very interesting to see kind of what it looks like on kind of our end on the contact centers end to try to navigate that for the selfish the selfish AI bot who's just trying to game the system because now you know kind of all bets are off. So that that will be extremely interesting in trying to think through some of the dynamics of that. So again, This I think was less about giving answers and more about some of the things that are coming.

13:20I think they're coming heavily for the retail industry, for the, internet, for kind of those FCC type companies that are in those type of industries. I don't know if banking and a lot of the healthcare will be there first. [pricing remark omitted] It can make 15 of those phone calls out for you instantly. I mean, Grokbot is doing that stuff kind of for us right now. So it's pretty crazy. So again, I wish I had more answers. I will be coming with answers as we go. Again, I think.

[Gap 14:10–14:42: no reliable transcription]

14:42, whole new world and it should be very interesting to see how we all kind of manage this this next even again 12 to 12 to 18 12 to 24 months

Your patients deserve agents who understand the stakes.

USA-based. HIPAA-compliant. AI call QA.

Book a Discovery Call
Or call: 877-766-7822