How AI is Changing the Sales Cycle | DJ Perry | TEC Services | Ep. 8

For Ep. 8 of our Applying AI series, DJ Perry of TEC Services Group joins Adam Parks and co-host Mike Walsh to discuss how artificial intelligence is reshaping sales across the receivables industry. 

Listen to Your Favorite Podcasts

Adam Parks (00:06)
Welcome to Applying AI, the podcast where we talk about the application of artificial intelligence to our everyday lives in regulated industries. Here with my season one co-host, Mr. Mike Walsh, and season one sponsored by EXL. Let's get started.

Today we've got DJ Perry. DJ is a guy that I've interviewed on the Receivables Podcast before, well-known sales folk from around the industry and a regular at industry conferences for the last
I don't know, 10, 15 years.

But DJ's got a lot of experience working across not only selling, but actually using artificial intelligence from a sales cycle perspective and selling artificial intelligence, which I think is a really interesting topic for us to cover here today. So DJ, thank you so much for joining us. We really appreciate you coming on and sharing your insights.

DJ Perry (00:56)
Thanks for having me, guys. Really looking forward to the conversation today.

Mike Walsh (00:59)
I'm gonna cut you off here, Adam, because so many of the greatest conversations in this industry that I've had are with sales guys, other sales guys, either late night or early morning, before everybody starts rolling into the exhibit hall. Hey, what'd you do? But there's so much information and and and so much of our industry is driven by growth, right?

Like, it is a race to growth. And that's why I really wanted DJ on and to talk about. Sales and AI affecting sales because it is, to me, the biggest topic there is. And I'm you know, as a sales guy that might be greedy, but I'm excited.

Adam Parks (01:33)
Well, as a reformed sales guy, I can tell you it's definitely the lifeblood of the business, right? Cash flow and being able to run our businesses and the sales cycle is changing when we're dealing with artificial intelligence. But before we get to that, I have plenty of questions for you, DJ. But for anyone who has not been as lucky as us to get to know you through the years, can you tell everyone a little about yourself and how you got to the seat that you're in today?

DJ Perry (01:56)
Yeah. So I started on the phone about fifteen years ago, like Adam said, with C B E, and worked my way into operations management. You know, a very large, sophisticated, high performing agency. So got to see a lot of amazing leaders, got to see a lot of things done the right way, which was really cool across all different debt types.

Then I started Locate Smarter, so started with data, which is kind of the foundation of AI, which is kind of interesting. So just having a little bit of that context. Then I did IT and cybersecurity managed services for a couple of years. So got to see how the cheese was made behind the scenes. That was fascinating, and then I went over to interactions, and then to AI, and then spent a few years on the dialer. So trying to hit all the different areas of technology and trying to stay on the forefront. It's been an interesting journey, but that's how I got here.

Adam Parks (02:43)
Well, it's really going to help you understand the criteria in which decisions are being made across each one of those tools and has probably given you a really good perspective as you've gone into selling artificial intelligence services tools. I mean, I think there's a couple of different directions we could take that. So, you know, look, both of you guys are actively engaged in selling artificial intelligence to the debt collection space.

Both of you have workers selling other products to the space. So you understand the true uniqueness of the debt collection industry when it comes to sales. So help walk me through that sales cycle that you're facing today. And how is it different selling artificial intelligence versus selling the other products and services that you've sold through the years?

DJ Perry (03:31)
I can jump in here, Mike, if you'd like.

Mike Walsh (03:32)
You go first. You're the guest. Go ahead. Yeah.

Adam Parks (03:33)
Kick it off for us, DJ.

DJ Perry (03:35)
Okay. Well it's it's it's been it's been fascinating. You know, as you guys probably know, or maybe some of you don't, you know, TEC acquired Latitude. Latitude is really an AI hub, if you will. And we work with a lot of partnerships through TEC Solutions programs. So we're seeing a lot of AI vendors and we're seeing a lot of non-AI vendors. So I would actually say that.

Conversely, we're seeing a huge acceleration with AI tools, but then it's slowing down a little bit on the non-AI-related tools. And it's fascinating. I mean, a real life example, we pulled an executive on a call, we're doing a demo of an AI tool, and on that call at the end, they say, Hey, can you be on site next week? I'll assemble the entire C-suite. This is something we need to work on now.

That's crazy. I mean, that's like that's not happening before, right? Like that's the dream outcome of like a first call or a demo, right? And that's what we're seeing with AI. There's so much excitement, there's so much buzz around it.

Mike Walsh (04:31)
I see the same thing, right? Like and it's funny I just EXL just brought another salesperson, Steve Taylor, to the space and you know, he's like, That's it? We're gonna go there now like one demo and I'm like, Yeah, yeah, that's fast, right? Like, I also think you are right; a lot of what we do is explaining how you can take the old way of working and adjust to the new way.

Right. Like there's a little bit of an education process a lot of times if it's an agency that's, you know, been doing it traditionally, like, hey, what do you need? What do I need to do? Who do I need on the call? Like sometimes we have a prep call before the demo to say, Okay, you might want to bring your compliance, your IT, and this person, your business suite, because they're gonna make decisions quickly about this. Right. Like, I'd love talking to ops people.

I love talking to data analysts because, like, the information they can get from these tools, the data back, is amazing. And a lot of our clients are like, Hey, can you give us this? Can you give us a call list based on behavior? Sure, we can do it. What do you want on basic? Right. Like, so it is a much more collaborative sale than, you know, I worked for three traditional agencies, right? Like it was a different world.

When I worked at True Accord, we had to explain, like, this is how your requirements should now be; we're not gonna make phone calls, right? Like we're we're gonna do these emails and text. So seeing how both clients will change and then, you know, if we're talking agencies at them, we're giving them a cheat sheet. Like this is now your capabilities, right? Like, with this tool, you are now fully loaded and ready to go. And it's kind of like an AI arms race, right? So we wanna take those salespeople and empower them to talk to their clients.

Adam Parks (06:11)
I think that's interesting. Now, with all of these new people in the room, whose budget are you working with now? Because from a technical perspective, it used to be okay, we're now talking about IT budgets or we're talking about ops budgets. You know, how is that investment being looked at by potential clients as they're evaluating and allocating budget for AI tools?

Mike Walsh (06:32)
Good one.

DJ Perry (06:33)
I'm seeing all the budgets. Like I'm seeing ownership and CEOs getting involved immediately because it's tapping into budgets kind of across the board. I've seen it's a little bit more complicated to make business cases now than it has been in the past, just because there are so many different areas that are involved. Scoping something out, making estimates on implementation time.

Getting it all to work together is very complex. It's fascinating. But yeah, I think it's hitting multiple budgets and just kind of being looked at from the very top down as a part of a biggest strategic initiative. That's what I've been seeing.

Mike Walsh (07:04)
I think you're right. Like, unless they have a digital transformation budget or AI budget, right? Like some companies will have that. It's usually creditors, or they have already assigned someone and given them the budget. And what's great about our space is it's a quick turnaround, right? Like collections. It's a great place to test 'cause it's the people who are not paying you if you're a creditor and it's the customers who don't pay.

So you know your liability is a lot a lot less and it's a great place to test like it's very measurable you know you know part of the process is that ROI measure right and you know what we do at EXL is we say this is going to be the reduction of your cost and this is going to be the increase in your lift and we'll put in right. So you know what you're gonna spend.

The reason you can do that is that AI can predict what you're gonna do. And it's pretty good at it, especially after years and years of doing you know, doing this. So that's a huge advantage in the sales cycle. And that's changed a lot of things. That's changed conversations for me and how because I think you're right, DJ. This is like an operational change and an IT change and it covers a lot of bases, right? Like it's even an HR change. Yeah.

Adam Parks (08:18)
And that's why I ask about the budget, because you're crossing across all these departments in all of these different allocations. You're trying to, you know, you're making an artificial intelligence spend. Does that mean that a portion of that falls under each one of those departments? And how are you or how are organizations looking at and measuring that mix?

DJ Perry (08:36)
Yeah. What, so I mean in the past we'd maybe see a VP of operations get excited about a tool. Maybe you're starting a sales cycle with someone, you know, well in advance. Now maybe that VP of operations, they have their CEO or COO or somebody right behind them pushing them. And maybe they're evaluating one tool, but they want to know what else. What else? What else?

And so maybe you know EXL can solve all their problems, maybe not. I don't Mike, are you running into places where they're asking what you can help with, but then it's bleeding over into areas that are maybe beyond what you can? Because it's like when the internet came out and they're like, what, how can we use the internet to improve our business? You know, and it's

like the same with AI. How can we use AI to improve our business, and they want to learn about all the different areas and use cases at the same time?

Mike Walsh (09:24)
That's a great question, DJ, 'cause there are calls I'm on, and I literally don't know what people are talking about. It goes really like it starts with collections, then it gets into customer service, and then it gets into modeling, then it gets into data analytics, then it gets into all sorts of things like just data warehousing and data lake building. Yes. So if you know, and we have these tools, but they're out of our space almost, right? Like sometimes they're in it, right? Like, you know, we're

We're talking to creditors about specific differences like healthcare, right? Like, well, our virtual agents are now doing insurance calls, right? Like, why do you have a human being calling an insurance company that's gonna take two things, like, just have bots do it. And then calling the customer to get their insurance because it was illegible, or they were passed out and they got up and left the hospital. Like all these types of little they're kind of our bit, they're ancillary things.

But they're very manual, and they're very expensive. So why not solve them with that?

Adam Parks (10:19)
Exception scenarios.

Mike Walsh (10:21)
It is a very easy leap. And those things we're doing more and more like re auto remarketing is another one that's popping up where people are like, Hey, can you help us with this? So yeah.

Adam Parks (10:32)
So once an organization makes the decision that we're gonna go forward, like we want to deploy XYZ tools. We're gonna bring it out into our organization. We're still going through a procurement process. And I know over the last 15 years, procurement documents have extended pretty significantly. You know, what's going on once they've made that decision? The CEO gives you the thumbs up, or whoever the approver is gives you that thumbs up. What's the next series of actions that's happening, and where do you see organizations falling off the rails in executing the decision that they've made?

Mike Walsh (11:09)
I think you have to guide, right? Like you can't let them fall off the rails. All right. Like, DJ, this must be complicated in your world too, because you have a SPOC, right? So you have all these things.

DJ Perry (11:19)
Yeah. Actually, they make you have skin in the game to make the implementation work, right? So you're invested immediately. May not all be on you, but

Mike Walsh (11:29)
No, but like you know, part of the process to make these things work is an implementation timeline and plan and then letting them know what we do and what they do, what they have to give us. Like if you don't give me a subdomain, I can't send emails, right? Like that's gonna happen, right? Like so and it's different, right?

Like agencies are used to like getting something and putting it in and going and then figuring out the bugs and what we do is you pre-solve it and and you test it and then the bugs are already out, right? Like, the thing works. It's different. But you have to say, look, I'm gonna need some time for your compliance team to review all these templates, all these emails, all this like pre-do it. And you know, sometimes Google doesn't like what they changed, right? And we have to come back and say, hey, Google doesn't like this. You gotta change this work. So there's a walkthrough period, but once you get on, I think it's not falling off, it's adding to it, right? Hey, can you handle the bankruptcy calls now? Like we're gonna go live. Like testing is over.

DJ Perry (12:32)
Yeah, so like AI urgency is n does not equal AI readiness. And I think that is really what we're seeing right now: there are CEOs that we've all talked to at the conferences that have been cutting edge for many years, and they're walking around with a smirk on their face because they saw this coming; they attended the trade shows, they were listening.

They're bought into change, and they're flying. And then we also have these companies now that are coming at us and they're like, AI's a thing, I guess, and we gotta figure out a way to lower expenses and go. And there's more pressure internally in those organizations to adopt faster. And then that's where you start seeing a lot more risk. So the implementation fails most often when, just to your point, Mike, the work wasn't done ahead of time.

Too many assumptions were made; people didn't check enough boxes; they didn't look under enough rocks; they didn't really take a close enough look at what needed to happen, and then things stall out, you know, roadblocks appear. And I think that's a big part of it. Like data integrity, data strategies, just resources and and really, really understanding exactly what it's gonna take. That's where I think I see it falling off the rails, so to speak, is just people aren't as ready as they probably could have been.

Adam Parks (13:48)
Their data wasn't as ready to be moved. They weren't as ready to analyze that volume of data. Like they just weren't prepared for what needed to come next. And I think that's one of those big challenges that we start seeing: those guys that are walking around with the smirk on their face have been preparing their data for years. They've been organizing it and then collecting it in that new format, knowing that someday I'll be able to use this in new ways. Not everybody was prepared to invest in the storage company.

Cost, the increased cost of insurance, right? All of the things that go along with maintaining more records. So, Mike, the question to you is where do you see like they've made the decision? Where are you seeing deployments go off the rails?

Mike Walsh (14:31)
That's a great question. 'Cause the th probably the most common is they didn't think of everybody who needs to be involved.

Adam Parks (14:40)
And it sounds like you're guiding them from day one on who should be involved.

Mike Walsh (14:44)
True, right? But that person might be on another project, right? Like they might be, let's say they're adding a payment processor and they're trying to add the AI because they really want to save money and cost. You'll see where they'll say, guys, we've got to delay this a couple of weeks. You know, it's not major.

Again, I think when you're choosing a vendor, make sure they hold your hand through the process, and they stay with you after it's done. I think.

That's a big part too, is once you go live once you test, we're we have live sessions with the testing. So they hear, you know, something's wrong or screwed up. They hear it, we hear it, we fix it. Right. And then a week later, you're now living with real debtors and you're working with them to say, Okay, these emails aren't good. Or these: you sent us fifty thousand accounts and only three have phone numbers.

Did you mean that? Like, you know, some of those are just simple, like they sent the wrong file, then they can do the stuff they meant to exclude, they sent little things like that, you gotta be aware of. And then you're reporting on now, you're reporting on delivery, and then you get into recoveries and walk in through all this new data they have. And now they can really break down their segmentation to micro segments. And you have to walk through that stuff.

Like we can't expect people who've never used an AI tool to magically wake up and know how to really get the max out of it. So part of what we do & DJ, I'm sure you guys have the same thing, is teach them how to use this thing to the max. And then sometimes it's baby steps because there's so much like Adam, your last question was like, where do you see this or does it go out? Yeah, then they start getting if you collaborate, you can start getting other use cases, even if it's not pure, pure collections.

But that saves you time, money, and just makes everything streamline because you want to build a partnership, right? Like it's just like everything I've ever sold. I've never sold anything to sell it and see ya. I'll talk to you now, at the next place, right? Like I've sold it to have a relationship and and have a partnership with my clients that, you know, you want to do more, right? Like you wanna,

Mike Walsh (16:51)
You wanna, and this technology is advancing. And it's incredible. Like we scrapped our voice from two years ago when I first started, and it's a JM now. It's amazing.

Adam Parks (17:03)
Let me ask you this from a sales perspective: who used to be the hardest person in the room to sell to, and how has that changed? Who's the new harder person to get through? Because it used to be finance and IT. IT doesn't have the resources, finance doesn't have the capital, but it sounds like the motivations within the organizations have changed. So have you seen that shift start to happen within the target?

DJ Perry (17:26)
I've seen some interesting shifts there where the CFO is all about the adoption because of the business case, because of the expense savings. and it's IT who's like, whoa, whoa, whoa, whoa, whoa, either we don't have enough resources or I think we can build this ourselves, or something to that extent. CFOs don't love that, by the way.

CFOs are like, Really, CTO? You sure you want to build this software and become, you know. So we've run into that where it's like, okay, CFO's like, please choose a vendor. But Mike, I don't know if you've

Mike Walsh (17:54)
I think it's changed too, like from three years ago to today, right? Like three years ago, compliance was scared of it. Now compliance is like, wait a minute, this thing doesn't mess up, let's go. Right? Like, and it is; I do get…

Adam Parks (18:06)
Well, there's validation in the marketplace now. They didn't have that a few years ago and you know, all you can fear is the unknown.

Mike Walsh (18:14)
Right, exactly. And you know, right, like, so they're kind of flipped totally. And now I do think DJ's build verse buy is still something. sometimes it's hosting, right? Like can you build this in our cloud? Or do you have to use the separate instance that we run? Like so there's advantages to that. So it's probably still a little bit of IT resources, but there's ways to cut that timeline down. Like, you know, we I I think IT wants, you know, they're technical people and they want to control it. But what they realize eventually is yes, they're gonna control this anyway. It's just gonna be a usual like for collection agencies on their own cloud, that's not in their environment, but is basically their environment, just an extension.

For creditors, yeah. Sometimes creditors have to build it within theirs, which is a challenge, right? Like it's not easy to do and it costs a lot of money. Because you have to get people you know, people who are now allowed to get into their system to build it and train their people on it, and it's different roles; they're gonna have to hire some people who know how to do this stuff. So there's a whole bullet. But I think the biggest change is that compliance people are now on your side.

Because AI doesn't have a bad day. It doesn't flip out on somebody, right? 'Cause you know, because their kid just called with a, and you know, crashed their car and now they're mad, and the next person feels their wrath. You know, like whatever that scenario is.

DJ Perry (19:42)
I was just gonna say some really easy ways to manage the compliance within these tools too, where you can update one thing and it'll push through and update, you know, fifteen or sixteen workflows or call flows all at one time. So it's like okay, compliance seems to feel a little bit more comfortable knowing that the updates that they need to make are actually being accelerated with AI. Yeah.

Mike Walsh (20:02)
Like real time almost, right? Like a day tops, yeah.

Adam Parks (20:05)
So now we start moving into the pilot's proof of concept, right? Now we're moving into that next stage of actually rolling one of these things out. And so this is kind of my first question for you guys, and it's a little bit of a loaded question here, but free pilot, paid pilot, no pilot, like what are we doing to test these tools? And what's the what have you seen in terms of an optimal way to give people an opportunity to feel some success?

But without becoming a free employee of that organization.

Mike Walsh (20:37)
I'll start, DJ, then maybe you can pile on with this loaded question that Adam so nicely sends us. I think we've done it for free in the past and what's happened is it doesn't get prioritized, right? Like within the organization. It gets delayed; it drags out. So we've kind of moved away. We do a test every time for the payment of a product, which is really what we're talking about.

Like I talk about. So we want you to test, get comfortable, and then sign the contract. So we wanna prove that it works. Our tests, our POCs, and our proof of concepts and our agreement, they flow together. Like so you don't have to go back to lawyers and do the whole thing. So it basically is we're gonna test for three months or six months, depending on what kind of portfolio and the volume. So that's a real test. Like there are real numbers for the AI to learn. Yeah.

Adam Parks (21:28)
Statistically valid sample.

Mike Walsh (21:31)
Exactly. That's the word I was searching for. Thank you, Doctor. but then you know, we put here's the measures of success that we're gonna give you, right? Like let's say it's twenty percent lift and thirty percent reduction in cost, like just throwing out a number. We hit those, we go, because most people want it to work and want to go, and we…once you build out the test, you really have the tools to start adding portfolios or clients. It's very simple. So I would say always test. and depending on what it is, though, there's certain things that are right, like if you're getting into a fraud system, I don't know how you do it. Or maybe it is just a one-year test and then you pull but like for us in collections, this is still collections, right? This is still a champion challenger versus business as usual. Or you can try another product versus ours, right? We do that all the time.

Adam Parks (22:21)
Yeah. And in a year, it's not a proof-of-concept test. Now you're testing the software, in my opinion, right? Like tests should be small, statistically valid approaches to verifying that a piece of technology, tool, or process change will have the impact on your business that you expect. And it sounds like your contract is if we hit these, now we start going forward. So if you've demonstrated that they're gonna get the value that you sold them.

Okay, now they're contractually committed. They're gonna start moving forward. I also agree on the free test because if you give it for free, nobody sees the value in it. It could be the most valuable thing in the world, but if you're giving it away for free or you're giving them that opportunity, I find that they don't commit the same level of resources, time, or attention.

DJ Perry (23:06)
I think it's also really dependent on the deployment and the level of customization required for the product. So if you have something that's just out of the box, no big deal, and it's a public cloud instance, you can spin it up and just say, here's your login for your sandbox, go nuts. Sure, free, lightly paid, whatever. That's a lot different than private cloud, fully custom tailored solution that requires thousands of hours to implement. Cause now you're like, to do the pilot or to do the test, you're basically deploying it. I mean, you're basically going through, right. Well well. I say basically because I've seen some like hybrid partial will kinda do implementations for the test, but then you're not getting the right No, that's exactly right. Yeah, it's like you're

Adam Parks (23:48)k
Have you ever seen that work?

Mike Walsh (23:49)
What are you testing? What are you testing? Right? Like you're testing an excuse.

Adam Parks (23:55)
Running concurrent systems to validate a test. You deploy it; you want to run concurrent systems for a period of time to verify, validate, keep yourself safe- a nice little safety
umbrella. That's a different story than hybrid deployment testing.

Mike Walsh (24:10)
Well you like, so we do modular deployments too, right?

Adam Parks (24:15)
But modular is different too because now you're going after individual use-case functionality in blocks in small chunks. That is a different approach to trying to roll out a hybrid. And I'm right. Well, but I think about it from DJ's perspective. The system of record can't do a free trial for three months. It's entirely too much of a I mean, you're talking about a year-long process to do it the right way. How could you ever accelerate that and offer it as a free service? I mean, just that the system conversion in and of itself is unique. There's no on/off switch; even if you're moving from the same system A to the same system B, it's just not that simple.

Mike Walsh (24:56)
But if I can get your email, you know, intelligent email and two-A SMS up and running in two months, you can test it, right? Like you can, it's two months, we're doing the heavy lifting and then a three month trial if you have a lot of accounts, six months over time if you wanna do less, right? So then it's easy. It depends on what DJ makes a valid point. Like at VoApps we used to give when I was there, shout out to Neal and and and those guys, as we used to give, hey, try if you'd. Like it was cool.

'Cause you could try for free, right? Like and you're like, well, this is cool. So, you know, it depends what it is, but I think what most people are looking at AI these days are, you know, kinda like the six use cases, Adam, that we've talked about a million times is like, you know, voice is very popular. You can deploy that pretty quickly and test it, right? Like you can test that. You can test two at the same time if you really want to. And you know, email text is pretty easy to deploy. Like those things.

Adam Parks (25:53)
So everything comes down to what is the actual product itself, right? How easy is it to deploy? So if we're breaking down the criteria of what I've heard from you guys, it sounds like the level of complexity, the level of investment, because if it's a couple of voicemail drops at whatever fraction of a dollar, like that's something that you can spend as an advertising cost. If it's

fifty percent of the sale of the software is in the deployment and the technical services. You just can't do the same things, which makes a lot of sense. But I was curious as to how you guys felt about those types of tool sets and maybe what experiences you've seen since you've worked across so many different, you know, kinds of organizations.

Mike Walsh (26:34)
When you get into a major, you know, CRMs with AI flows, could you know you gotta prep your client for like I hear nightmare cases of like, hey, we didn't know this thing was gonna take six months to deploy. Like, you know, some people come back to us, right? Like maybe they found a cheaper solution and like this was gonna be great and it didn't work, right? Like you gotta make sure you map out. Part of if you're in this business selling AI, you have to map out with your client. This is how long it's gonna take. I always build in extra time because clients, something pops up right like and is gonna go

Adam Parks (27:01)
Smart. Underperform, overdeliver.

Mike Walsh (27:05)
Yeah and just be like and people it's great what it's what's different is I'll get an owner of an agency call me and say are these guys dragging their feet like about their departments right like and I'm like no no no no no like you know like it's usually yeah they're excited about it so usually you because people see the value

Right, like it's it's you know, I'm sure when people go.

Adam Parks (27:27)
Well, so now that we're talking about the artificial intelligence tools themselves and we've kind of gone through the sales cycle. Now we've talked about it being deployed. I'm curious as sales professionals, how are you using AI today in your day-to-day workcases to support your sales efforts on behalf of AI organizations? And DJ, starting with you.

What's your tool of choice? What's your tool of choice?

DJ Perry (27:56)
So we use Enterprise Claude, and that's my tool of choice when I work. I think it's phenomenal with spreadsheets and analyzing, you know, pricing calculations and business cases, and a lot of content creation. Personally I love Chat GPT, so I might use that on the side if there's any non PII specific stuff that I can, you know, feather into that and maybe get some ad advice here and there.

But those are the two models predominantly that I've used. I know there are others like Perplexity; people speak highly of that. I have not explored it. Yeah.

Mike Walsh (28:30)
If you insult Perplexity, Adam will weep. He will weep on this podcast.

Adam Parks (28:37)
Hey, look, there's only one model where I can run all of the models, but that's just a different story.

DJ Perry (28:41)
Yeah, I'm gonna download it right after the podcast.

Mike Walsh (28:45)
I'm Claude as well. We have an enterprise solution. I think it's great. And then if I'm trying to figure something out quickly and ask my phone while walking in an airport, I'll ask Gemini, right? Like that's probably my personal one. But I do think Claude is amazing. Like slide decks, we and then we have tools internally.

Like if you think about where we started, you know, in sales, DJ, like two guys who introduced me to you, Nick and Mikey D, they taught me years ago when I was working with them how to use AI to shrink my emails. I'm like, what are you guys talking about? Like, I had no clue as the old man. And like, you know, if I think about it, that probably started, and then I used RFP 360, which was a great product, which really wasn't AI; it's probably machine learning to fill out RFPs. And I love that product, right?

So those two things, and I forget what the little email add-on was. I think Copilot bought it and ate it. So I used that.

But those two little things save me so much time. And I think salespeople will always look at the most valuable thing they have is time, right? Like so there's only so much they can do in a day. So they'll look to cheat time every which way, and so many people like it.

Again, kind of how we got started being friends and talking is through other salespeople, and we always have these chats, and it's just great. Like, you learn these little tricks. So I think sales has been leading the AI push for companies. And if you don't think your salespeople are using AI, you're wrong. They're using AI. Like, they're way wrong. Right. Like they're using it. If they're not allowed to use it, just buy an enterprise solution because they're using it. They have to be competitive, I think. Like for research, it's so much easier now than it used to be.

Adam Parks (30:32)
Well, talk me through some of those use cases for leveraging these tools in the sales cycle.

DJ Perry (30:38)
Yeah, I well, I was gonna say, you know, I don't know a single successful enterprise account executive that doesn't like the dream of hiring an assistant, right? And then you get AI plugged into this, and all of a sudden they kind of have an assistant. Like it started off, yeah.

Mike Walsh (30:52)
Yeah, sales ops. Yeah.

DJ Perry (30:54)
It started off with making the email more concise and helping kind of shape when that's cool. Like, that definitely saves time. But getting into what you're just asking now, Adam, you know, a colleague of mine built a bot that can go out and research an account that he's prospecting, taking all the products and solutions and value ads that we offer into consideration and feed information into it to to say, hey, try this angle.

You might be able to get some interest generated by doing this. And then, you know, going into prep calls, having all that background information, having it review all of your notes, just the AI note taker on calls working so well to be able to reference and pull back up and ask questions is phenomenal. And then I mean, Mike, you brought up RFPs, man. I am three sixty, whether it's Cloud Direct, Responsive, whatever the tool is. What are you seeing with that? 'Cause I mean I think that's like the best use case for AI on the client side and the seller side.

Mike Walsh (31:48)
Yeah, and I think I like it, whatever tool it is, right? Like I think EXL has if you're in sales at EXL, you hit you drop down your claw and say what is your role? And some of those things are amazing that they have pre-built. So I don't have to build. They're already built. I can just use them. The scrapers, like our marketing team too, have a lot of that stuff. So that's you know, if you have a larger sales force or even for reporting, it's great.

Like we do a lot of work with Salesforce. So what they're coming out with, we just released some of those. I mean, it's just so much easier, right? And saving time. I've what is the even for expenses, there's some tools that are I'm jealous we don't have them yet.

Adam Parks (32:28)
Are you querying against your CRM? Like can you query against your CRMs and poll reports and you know, not spend hours building and trying to understand? I haven't used Salesforce at all.

Mike Walsh (32:34)
We're on Salesforce.

DJ Perry (32:38)
Yeah. Well it's the same, it's the same AI readiness problem on the vendor side as it is the client side. You were either documenting and filling in the data that you now hope that you can go query and analyze, or you weren't. Right. And so, I mean I've seen it successful, being able to go in and get insights on pipeline and what deals have been out there maybe a little bit longer, who needs attention, updating next steps, updating notes. There's definitely, I think, ways that we can accelerate it and I think it does come down to how well you were prepped and how well you have built.

Mike Walsh (33:11)
And what you tracked, right? Like, so much of it is about controlling what you track. Like, tracking too much is just wasting space and computing power. Really think about what's valuable, what's not. Like some of the cool things like Salesforce, I think I think it's Salesforce, I don't know if it's our internal tool. It will tell you; it will go to their link.

Let's say Adam Parks is my prospect. It will go to your LinkedIn and summarize who you are for me. Like where you went to college, like all this stuff. I don't have to go look. Like it's pretty cool. Like, that just saves time. And that's already there. It's connected to your account. It's cool stuff. So the tools are

Adam Parks (33:48)
Have you guys started building any automations, like, you're both using Claude? Have you built any Claude skills to, for example, take that sales transcript and update the sales call transcript, update your CRM, and prepare a response and a scheduling or whatever? But have you guys looked at any of those types of automations, or is that already more being driven from your CRM platform?

DJ Perry (34:10)
We're working on it. I mean, we're working on different things. Trying to get everything to work together again starts to become a little bit of a challenge. Like, in a perfect world, it would read the email and it would, you know, just get availability. Like, I know a lot of people use Calendly links. I just am, and maybe I'm unique. I and and and there's a good use case to like you but whenever I get asked to fill one out, I'm like, You're asking me for the meeting and I gotta go fill out your link.

Mike Walsh (34:37)
I love it, DJ.

Adam Parks (34:37)
Hey, think about it however you want, but I went from we did an analysis of my email and it was like it was like thirty five percent of my communications were about scheduling a meeting.

DJ Perry (34:47)
Yes, that's my point. And so I'm like, there's gotta be a high open. Yeah. No, it's just so

Adam Parks (34:49)
Yeah, and now it's like three percent. Now it's just Mike.

Mike Walsh (34:53)
No, but Right, 'cause I'm not gonna do it. I'm with DJ; I'm not gonna do it.

DJ Perry (34:58)
All you did to be like, give me three time slots this week. Like, give me three time slots for basketball.

Adam Parks (35:01)
But it's also about how you frame it, right? Send me your availability or you can check mine right here on Calendly. I try to take an approach where it's, you know, send me your availability. We can do the dance if you want.

DJ Perry (35:12)
Right. It's a great tool. It definitely saves a ton of time. And the second you click the link, you're like, this is way easier. I can just pick the time I'm available. Like me, I love the tool, but there is just something about it. So I'm like, well, if I could get Claude or whoever to just give me three time slots, bullet it out in bullet points. I can slap it in an email. It looks like I did it the old school way, but I didn't. Like that's, there you go. See, Calendly's all about I'm not anti-Calendly. I'm not.

Adam Parks (35:33)
I hate to say it this way, but you can do that in Calendly too.

Mike Walsh (35:40)
No, no, me neither, but I kinda am, right? Like But…

Adam Parks (35:45)
Me too, but it's saving so much time that I'm now committed.

DJ Perry (35:47)
It's yeah. I get it. But you plug all that stuff in; it does help. You can get it all to work together. You get your next steps update. Like, yeah, I think there's a a world where it all works really smoothly. I was actively working on it, so maybe we'll have to do a follow-up podcast.

Mike Walsh (36:00)
Yeah, I think some of the stuff gets to the point where like too much I don't mind the automation in the background, but I will never automate the outreach to prospects and clients. Like I just won't do it. I don't care if it saves me time or not. Like, I'll take the time doing it because…

Adam Parks (36:17)
I concur, but that's the time and place for that personal touch. That's where the relationship is built. That's where the shared experience happens. I agree with that wholeheartedly.

DJ Perry (36:26)
I have a funny story today. I have a funny story. I got a prospecting email from a lead gen company. And it's D capital D little J. It's just a little pet peeve of mine. It's not a lowercase J anywhere, right? So if anybody did any research, they would have seen how to write it. So I'm like, you know what? I was a lead development rep for a few years. I'm gonna just help this guy out. So I started going through, pointing out a few things, and there was more than just my name.

Anyway, I got to the point where I'm looking up his profile on LinkedIn. It was a bot that was prospecting me. There was no feedback to give to this dude. I'm like, this is I'm I've just now wasted my time trying to help a bot if I guess. Yep.

Mike Walsh (37:06)
A bot that was prospecting me, yeah, yeah, trying to help it.

Adam Parks (37:09)
Yeah. You're gonna help improve their Claude skill.

Mike Walsh (37:13)
Those are like how to lose a client in one second, right? Like send one of those, right? Like and and I was just at our RMAI exec. I don't know when we're gonna release this, but I was just there a couple of weeks ago.

And so many people came up to me like, Walsh, can you tell these other AI companies how to send a sales email like you know, or this isn't the time or place for feature selling. You know, it is our industry; there's like a code of conduct as salespeople that you learn. I think there's been a lot of new players that come in and just think they have the greatest product, and maybe it hey, maybe it is. I've never seen it, but simmer down and automating blasts to

Everyone at a conference, if there are 20 people from the same organization and they get the same email, you look like an idiot. You just do. Right? Like stop. And

Adam Parks (38:02)
Well, that's not prospecting; that's blasting, and the marketing team and the sales team should not be doing

Mike Walsh (38:06)
That's laziness.

Mike Walsh (38:08)
Yeah. It's just

DJ Perry (38:09)
Perceived effort, right? Like, if you do a handwritten letter nowadays, it goes over super well. 'Cause it took a lot of time and effort. I mean it, I think that's a big problem with Mike.

Mike Walsh (38:16)
No one can read my writing though; that's the problem.

Adam Parks (38:21)
I believe there is a third-party service for that too, Mike.

Mike Walsh (38:23)
Really, it's my wife; she can write. But I can't.

DJ Perry (38:25)
Really? No, that's awesome.

Adam Parks (38:30)
No, there's a couple of companies out there doing some interesting things to bring that personal touch back into the sales cycle. And I know as silly as it is, I've always used postcards and handwritten cards after conferences for years. And I still find them in people's offices when I show up, you know, three, four years later and see that they've got a card on a credenza or whatever. Because I think that personal touch is important. So you bring up a really good point. I do think there are some interesting ways to use the automations.

To be able to ingest information, document management, storage, sticking things into folders, organizing your discussions and preparing even before I get on the call, give me a quick summary, make sure I'm prepared to, you know, I've done my research and I'm walking into this prepared. And I think this goes back to the same ways that I look at artificial intelligence from a debt collection standpoint. The more that you're accentuating that live agent, you're feeding them information, you're giving them the right tools to be successful, the more successful they can be.

And I think that sales are kind of the same way. The more that we empower those sales folks and we give them the tools that empower what they're doing, not replace the actions that add value, is where we're going to continue to see sales change. And it does sound like there's some pretty big changes in how people are buying artificial intelligence tools versus how they have purchased in the past. And I would love to do a follow-up next year and learn more about how organizations are.

As they're deploying these tools and this technology and they're starting to find this benefit, what happens next? Like, where do they start to move into new and other products? Are we starting to compare the best use case against each other? Are we continuing to push the boundaries once we've crossed the finish line of a deployment? And how are we continuing to work with those clients to grow the opportunity that they've purchased from us?

DJ Perry (40:19)
There are so many things we can talk about at this point that we can go so many different directions right now, which I absolutely love. Yeah, I think there's a filter in our space right now. And if you're enabling and allowing your clients to make the choices that they want to choose and work with the vendors and the products that they wanna work with, if you're making that easy, you're gonna pass. If you're not, and we all know who those are, you're not gonna make it.

So I think there's some things that are gonna be really interesting to see unfold over the next, you know, call it one to five years. The landscape's gonna change quickly and there's some new folks coming in that are making it real easy to buy and it's going real well. And I think it's taking a lot of the calls by more legacy vendors by surprise. So it's exciting. We're gonna see some shake-up in a good way.

Adam Parks (41:07)
Well, sounds like we've got a lot more to talk about in a future episode. DJ, really appreciate you coming on and sharing your insights here with us today. This has been a great discussion, and sales folks are really a lot of the lifeblood of this business. And being able to look at artificial intelligence and how it's being applied from their perspective has been insightful. Mike, anything you'd like to add?

Mike Walsh (41:30)
No, I think DJ made a very good point. The industry is changing, and I think you're right, Adam. Like salespeople are changing in this industry. What their focus is has to change. You can't say Come see our call center anymore. It's if we have a really great training program. Like what's important now has changed because of technology and finding ways to articulate it and and educate your prospects and I think a lot of things too, like DJ says like partnerships are way bigger than they ever were. Right. Like you used to worry about yourself and that's it. Now half of what I do at a conference is talking about possible partnerships.

Mike Walsh (42:10)
And I think that's because if two techs fit together very well and complement each other, let's get them free integrated so it's easier for that c customer. So I think we're gonna see a lot more of that too.

in the next couple of years. But we're seeing the wave hit in many, many ways. And like it's been fun and this was a great conversation. Again.

DJ Perry (42:32)
Yeah.

Mike Walsh (42:33)
I was excited about this one, Adam. You know that. I was pestering you to have this. And I think we should come back in like six months and do it again and and just because I think that's the speed. We could wait a year, but I think it's a totally different universe.

Adam Parks (42:48)
It's happening really quickly, but I love getting that insight after you've been to a couple of conferences. You just came back from ACA, you just came back from the summit, you just had all these conversations fresh, and it feels like the perfect time to have that.

Adam Parks (43:00)
I really do appreciate every go ahead, DJ.

DJ Perry (43:03)
I was gonna say, I feel like we could talk for another three hours. We literally just got started. Like there's so much more to talk about. But I did want to make sure I ended on a note that I think is really crucial. So I think everything's changing with AI, but I think the things that make a good salesperson great and what makes us successful have not changed. Like, people are still people, and you still have to earn trust, and you still have to build relationships.

And at the end of the day, it's still a human being going, Yes, I'm choosing to partner with you long term. So I think consultative sales approaches and those that focus on the relationships are still going to win out, and there's no AI tool that can replace it.

Mike Walsh (43:41)
Hundred percent.

Adam Parks (43:41)
I agree. I don't think we're getting replaced any time soon, but we're gonna sure have some fun talking about it in the short term.

Well, we do appreciate everybody's attention today. Thank you for listening to Applying AI, where we explore how to make artificial intelligence work in the real world of regulated industries. Subscribe to the show on your favorite podcast platform or YouTube, and you can find more insights at receivablesinfo.com. We'll see you all again soon.

Why the AI Sales Cycle in Debt Collection Feels Different

One demo. Then a request to come onsite the following week with the entire C-suite in the room.

That is not how most enterprise technology sales traditionally move. But according to DJ Perry, it is an example of what can happen when artificial intelligence enters the conversation.

The AI sales cycle in debt collection is developing at its own rhythm, and this changing dynamic sits at the center of this episode of the Applying AI Podcast. Host Adam Parks and co-host Mike Walsh from EXL are joined by DJ Perry of TEC Services Group to discuss what it actually takes to sell, buy, test, and deploy AI in receivables management.

Faster AI Sales Require More People at the Table 

Perry describes seeing “a huge acceleration with AI tools,” including the example of an executive responding to a demonstration by asking the vendor to meet with the C-suite the following week.

Walsh sees another change: the sale has become more collaborative. Before a demonstration, his team may help an organization determine whether compliance, IT, operations, data, and other decision-makers should already be in the room.

That matters because excitement does not eliminate implementation complexity. If anything, faster interest can expose organizations that are enthusiastic about AI before their data, people, processes, or expectations are ready for it.

The episode follows that tension from the first sales conversation through implementation and then turns the lens around: How are the salespeople selling these tools using AI themselves?

The answer reveals an interesting boundary. AI can research, summarize, analyze, organize, and save time. But should it be trusted to build the relationship too?

AI Buying Decisions Are Moving Up the Organization

Perry is also noticing a different mix of decision-makers.

“I'm seeing all the budgets. Like I'm seeing ownership and CEOs getting involved immediately because it's tapping into budgets kind of across the board.”

This creates a more complicated business case.

An AI deployment may affect operations, technology, staffing, compliance, analytics, and other areas simultaneously. Walsh notes that collections can offer an especially measurable environment for testing because organizations can evaluate outcomes such as cost reduction and lift.

The discussion also highlights a shift in who may support or challenge the purchase. CFOs are interested in the expense-saving business case while IT raises questions about resources, control, or whether the organization should build the technology itself.

Meanwhile, compliance is becoming more comfortable with AI than it was several years ago.

The takeaway is straightforward: selling AI may require selling the operational change around the technology, not merely the technology itself.

AI Proof of Concept Strategy Needs a Definition of Success

Once interest becomes intent, another question appears: How should organizations test AI?

Meaningful testing starts with proving that the technology works. Effective POCs require sufficient volume, agreed measures of success, and a clear path from testing to deployment. 

Not every technology can be tested the same way. An out-of-the-box tool running in a public cloud environment is very different from a heavily customized private-cloud deployment requiring significant implementation work.

This suggests that an AI proof of concept strategy should reflect the product itself.

A lightweight communication tool may lend itself to a relatively quick test. A major CRM or system-of-record implementation may not. In those situations, pretending every product can fit into the same “free trial” model can create unrealistic expectations before implementation even starts.

AI Sales Productivity is Growing Behind the Scenes

Instead of asking how organizations buy AI, Parks asks Perry and Walsh how they use it as sales professionals.

Perry's answer quickly moves beyond content generation.

“I don't know a single successful enterprise account executive that doesn't like the dream of hiring an assistant, right? And then you get AI plugged into this and all of a sudden they kind of have an assistant.”

That “assistant” is already taking on a wide range of behind-the-scenes work. The conversation touches on account research, prospecting preparation, reviewing notes, RFP support, spreadsheet analysis, pricing calculations, business cases, CRM analysis, pipeline insights, next-step updates, and meeting preparation.

The common denominator is time. 

Used well, AI sales automation can reduce the administrative work surrounding a relationship. A salesperson can arrive at a meeting better researched, find information faster, and spend less time manually organizing what already exists.

But that only works when the underlying information is useful. CRM intelligence creates the same readiness challenge seen on the buyer side. If a sales organization has not documented and maintained its data, there is less for AI to analyze.

Human-Centered AI Sales Still Depend on Trust

There is one area where Walsh draws a hard line – “I will never automate the outreach to prospects and clients.”

After receiving a poorly personalized prospecting email, he started looking into the sender and discovered he had been responding to automated bot outreach. Technology has created activity. It had not created a relationship.

The importance of “perceived effort,” noting how something as simple as a handwritten letter can stand out precisely because the recipient recognizes that effort went into it. This becomes the broader lesson of the episode.

AI can help salespeople prepare. It can organize information. It can surface opportunities. It can make research faster. What it should not automatically do is remove the human action that creates value.

AI Sales Cycle in Debt Collection: Actionable Tips

  • For receivables organizations buying or selling AI, the conversation suggests a practical checklist:
  • Bring operations, IT, compliance, finance, and other relevant stakeholders into the discussion early.
  • Assess AI implementation readiness before allowing urgency to dictate the timeline.
  • Map client and vendor responsibilities before deployment starts.
  • Define measurable success criteria before beginning an AI pilot.
  • Match the POC structure to the complexity of the technology being tested.
  • Use AI to strengthen sales research, meeting preparation, CRM analysis, and administrative work.
  • Keep CRM and pipeline data useful if you expect AI to generate meaningful insights from it.
  • Protect the human touch in prospect and client communication instead of automating every interaction.

The AI Sales Advantage May Still Be Human

The AI sales cycle in debt collection is changing quickly. Buyers are moving differently. Decision-makers are changing. Sales teams have new tools. And implementation readiness is becoming part of the sales conversation itself.

But Perry closes the episode with a useful counterweight: “People are still people, and you still have to earn trust, and you still have to build relationships.”

That may be the most practical AI sales strategy in the entire conversation.

For more insights into how AI is changing all aspects of the receivables industry, explore more episodes on Applying AI, and visit Receivables Podcast for more conversations with receivables industry leaders.

Key Moments from This  Episode

00:00 – Introduction to DJ Perry and Mike Walsh
03:35 – How AI is accelerating the sales cycle
17:26 – How AI is changing decision makers and compliance conversations
27:56 – How sales professionals are using AI today
30:38 – AI for prospect research, RFPs, and sales preparation
32:38 – AI readiness, CRM data, and pipeline analysis
43:03 – Why human relationships still matter in AI sales

Frequently Asked Questions

Q1: How is the AI sales cycle changing in debt collection?
A: DJ Perry of TEC Services Group and Mike Walsh of EXL Service Holdings explain how the AI sales cycle is bringing executives into buying decisions earlier, accelerating evaluations, and requiring greater coordination across operations, IT, compliance, finance, and other teams.

Q2: What does AI readiness mean for receivables organizations?
A:DJ Perry and Mike Walsh discuss why urgency alone does not equal AI readiness. Successful implementation can depend on data integrity, internal resources, stakeholder involvement, realistic deployment planning, and clearly defined success measures.

Q3: Can AI sales automation replace human sales relationships?
A:The discussion suggests that AI sales automation works best when it supports research, preparation, analysis, and administrative work. The human side of the AI sales cycle, including trust, personal communication, and consultative relationships, remains critical.

About Company

TEC Services Group

TEC Services Group is a technology and professional services firm serving the credit and collections industry. Its solutions span collections technology, managed services, data management, implementation, consulting, and automation, helping organizations improve recovery performance, efficiency, and compliance.

About Guest

DJ Perry

DJ Perry is a Solutions Account Executive at TEC Services Group with extensive experience in collections and accounts receivable management. His background spans operations management, data and analytics, technology partnerships, IT, cybersecurity, and consultative sales within financial services and collections.