Exploring the Power of AI in Payroll with Special Guest Amy Travis

[00:00:09] Welcome everyone to another episode of HR and Payroll 2.0. I'm Pete Tiliakis and as always, I'm joined by the legendary Julie Fernandez. Welcome Julie. Julie. Thanks Pete. Great to be here and today I'm super excited because we have a practitioner friend to introduce, Amy Travis. So Amy, do you want to introduce yourself and like can't wait to talk to you about AI stuff.

[00:00:32] Amy Travis Yeah. So I'm Amy Travis and I am based in Atlanta, Georgia. I have lived here my whole life. I'm involved in the local payroll chapter, actually president elect, have been doing payroll for about 20 years. I came from the bottom, rose to the top.

[00:00:50] Amy Travis And have just currently I'm in a senior manager role at West Rock, Smurf at West Rock, and just enjoying meeting all of the people there. I just joined that company about two months ago. So as we all know in the payroll world, two months feels like two years. So it's been a lot of fun dissecting everything and seeing where I can help out. That's my favorite.

[00:01:15] Steve Yeah. Fellow ATLian. Glad to have you. And shout out to Ainsley and her team. It's a great chapter, right? Amy Travis Yes. Steve So yeah. Excited for you. Take the reins. Amy Travis Yeah, for sure. Amy Travis A little nervous, but it's okay. Amy Travis Yeah. Amy Travis Fun. Good fun. And that Atlanta group is really kicking it. I mean, they're super active. Steve They have all the awards. They have all the accolades to prove it too. Amy Travis It makes me want to be an Atlantean, but alas. Steve An ATLian. You can come. You can be on. You're here often. You have clients.

[00:01:45] Steve An ATLian Yeah. Thank you. Steve An ATLian We'll accept. We'll take you. Amy Travis Don't call it hotlanta. That's not a thing. Amy Travis Okay. That's not a thing. Amy Travis Good to know. Amy Travis That was the 90s. Amy Travis Yeah. Amy Travis Good to know. Yeah. There's probably more than a few jokes by now about that, right? Amy Travis The big peach. Amy Travis Yeah. Amy Travis Hey, Amy, we love to start with how you got into payroll and ultimately then why you stay, but everybody seems to have some kind of a screwy path to getting here. So what's your- Amy Travis It's almost like we, uh, it's almost like you need to start asking people,

[00:02:14] what's wrong with you that you, you know what I mean? Like, it starts in some, it starts in some, you know, not directed way that they landed there. So we'd love to hear it. Amy Travis Yeah. Tell us. Amy Travis Yeah. So my story starts at a Sonny's barbecue on Cheshire Bridge Road in Atlanta. Amy Travis Okay. Amy Travis Okay. Amy Travis There you got food. Amy Travis So I was working at the restaurant. One of the district managers comes up to me and Amy Travis says, what would you do if you never had to butter another piece of garlic bread in your life? Amy Travis And I was like, sign me up.

[00:02:44] Amy Travis So what had happened is that our corporate office, Amy Travis So I had worked for a group that had owned Golden Corral and Sonny's barbecue at the time. And they had a wage audit where they weren't paying overtime on their tips. They weren't including that in the regular rate of pay. So I actually came in Amy Travis And started keying for them at this group, manually keying and Microsoft Dynamics. Amy Travis This is like 2004.

[00:03:14] Amy Travis Wow. Amy Travis And so I guess they liked my data entry skills. Amy Travis And they were like, hey, so do you guys, Amy Travis Do you want to join the payroll team? And so that's kind of where it started. Amy Travis And so I, you know, I went from doing just data entry back in the day Amy Travis And I had a really, really good run at another company where I was, Amy Travis Went all the way up to a payroll director.

[00:03:41] Amy Travis And so I guess during that time, I just realized that Amy Travis The way my mind works and how I can retain information and compliance and tax and, Amy Travis Do you know, do the research and the digging and stuff like that? Amy Travis It just kind of, it kept me because there's, Amy Travis There's just never a dull moment. Amy Travis And I, I, it's, it's funny, I thrive in chaos. Amy Travis And I feel like, you know, we don't want payroll to be chaos, Amy Travis But it is.

[00:04:09] Amy Travis And so that's kind of, that's kind of what, Amy Travis What made me stay. Amy Travis But it's, it's been, it's been a lot. Amy Travis And, you know, like I mentioned the chapter before, Amy Travis Getting involved with them, getting my CPP. Amy Travis It just, it reiterated the love for payroll and now giving back to the community. Amy Travis Once you're in, you can't get out. Amy Travis That's fair. Amy Travis One of my favorite things is, is just pouring what I know into other people.

[00:04:36] Amy Travis It's a very common answer though, about the, about the challenges and the problem solving and all of that. Amy Travis I think it keeps people stimulated, right? Amy Travis It's not, it is the same every day, but it's not the same thing every day. Amy Travis So yeah, well, good. Amy Travis Good to have you. Amy Travis We appreciate your service. Amy Travis Having chaos thrust upon you is not something everybody enjoys. Amy Travis Yeah. Amy Travis Yeah. Amy Travis Or sticks around for the abuse. Amy Travis Indeed, indeed. Amy Travis Well, Amy, we had a chance to have some conversation the other day and I just really

[00:05:06] thoroughly enjoyed talking to you about some of the ways that you've been incorporating AI into your, your daily productivity, right? Amy Travis It wasn't like a certain payroll, you know, use of it or something else. Amy Travis But, but it's clear, it was really clear, easily clear to me that you've kind of embraced it and been messing around with it. Amy Travis And there's so much for me and for others to learn about that. Amy Travis So I wanted to give that a little bit of a dig and just kind of see,

[00:05:35] you know, like how, how, where did that come from? Amy Travis How did you start? What, what first tickled your fancy or got you interested in, in checking something out? Amy Travis Yeah, so there, I think this is about a couple years ago or so. Amy Travis I'm, you know, had found that people were doing these, like it wasn't just ChatGPT, people were doing these off apps, where you know, you could use the service, but in a different way.

[00:06:05] Amy Travis And there was someone in Texas that had to develop an application where you could actually just text to the AI server. Amy Travis And so that's where it kind of started in my, my personal life, right? So I would be texting to what was my personal assistant, help me plan the menu for next week. I have four kids, full time job. Amy Travis So if I can say, hey, my kids like this, blah,

[00:06:30] blah, blah. And I just was so interested in the fact that people could chat with these bots and get it to a level where the bot knew them, and what their needs were, and what they needed. So where like, I could explain like, hey, my kids like this, my kids don't like this. Amy Travis And so that's where I started,

[00:06:52] Amy Travis And so that's when Microsoft Copilot was coming out and trying to get integrated in that. Amy Travis And so that's when Microsoft Copilot was coming out and trying to get integrated in that. Amy Travis And so it really piqued my interest on the data integrity, making sure that you have a license, making sure you're not putting hot data into just like a regular web server. And so that's something that I always kind of preach, like if you're on Copilot,

[00:07:21] make sure that you're on work and not web. So doing the Copilot really helped me get into like, oh, this is what I can do in payroll. Amy Travis I just want to touch on that just a little bit, because I think a lot of people recognize that you don't necessarily want to be on, you know, wide open when you start messing around with, with AI. But, but how do you know if you're on a work,

[00:07:48] you know, something that is commercially available? Yeah. Or if you're just like out there in the wide open, how, how did you even find out or know? You have to ask? Amy Travis Yeah. Amy Travis You know, I would have never played with it on a work level, but I was with a company that had actually built out their own chat bot. And so we knew it was data

[00:08:12] integrity, and we knew everything was, was good. So that one was outside of Microsoft Copilot. They had actually built their own. And what was really cool about that was that, you know, there were, there were company specific questions that you could ask and get answers. And even today with the company that I'm at, I know that we had the conversation where I was like, wait, where's my AI?

[00:08:37] But just through digging and finding that there was a copilot, like a copilot pilot, I was able to, to like very quickly, I've only been with the company two months. And I'm, I think probably in my second week, I was, I was dying. So I, I found that there was a pilot. And I think that that's one of the biggest, probably things that I can say to anybody, is if you're with

[00:09:04] a company, especially a big company, and it could be a small company. If you don't have access to AI yet, or you don't have copilot, or you don't know if it's out there, do some digging, do some research, somebody in IT can, can help you. And you never know, like they could be doing a pilot program. And in my case, they were so excited, because I was in the HR space, that they had some upcoming series and stuff like that, where they wanted my input. Yeah, many of the, I've had a lot of conversations

[00:09:32] with executives about innovation over, over this, not just this podcast, but you know, the Daily Pay, the Source podcast, and obviously in my travels. And one of the things I find is that a lot of, a lot of eight, especially in bigger companies, the HR organizations and the finance organizations are getting the opportunity, I think, a better opportunity to have the organization had already done some front end work, whether it was a chat bot or something on their product side. And those learnings were now being sort of transferred versus them being the tip of the

[00:10:01] spear, which I think can then be, you know, all the pressures on right where now it's like, well, we've learned some things, we've made some investments, let us see if we can help you and your, you know, your team. So it's good that you have that forward innovation available. Yeah. And that's where the agents come in where Not everybody has it. You can, you can start building an agent and have your policies uploaded. This is, I mean, it's really helpful work. Just, you know, taking the simple, you know, FAQ items and

[00:10:31] throwing them into, into that bot for the employees too. Hey, there are payroll pros, leaders, and stakeholders. Pete Teliakas here with some exciting news. The second annual payroll profession confidence index survey is now live. Last year, the PPCI reached and amplified the voices of payroll experts from 28 countries across five continents. And this year we're taking it to a whole other level.

[00:10:58] The PPCI is groundbreaking research designed by payroll for payroll and aims to share your sentiment, impact and insights to help shape the future of the critical profession. The PPCI is a hundred percent unsponsored, completely anonymous and takes less than 10 minutes and is open now through March 30th. So don't miss this chance to impact the future of the payroll profession. Head over to payrollinfluences.com forward slash PPCI today to join the research.

[00:11:28] And please, please be sure to share this with your network. Let's make sure every payroll voice is heard. Results drop May 2025 at payrollinfluences.com. Any sort of key use cases that like right away you were sort of identifying or how, yeah, maybe what's the path that you went to identifying those in the first place and then sort of how you rolled this out? Yeah. So with my last company, it was a lot more. And so I say that because I'll probably reference

[00:11:57] that more because that was, you know, I was there for, for a couple of years and I'm very heavily involved in trying to get this rolled out and worked with the HRIS team. And so that was, that was our big project was getting these agents set up so that we could have that self-service. Because we're in the age of self-service, right? Like we want the employees to go out,

[00:12:23] view their own paychecks, update their own taxes, and why not also give them an opportunity to comb through the policies in an efficient and quick manner. And so just uploading those policies, uploading sometimes SOPs, we would, we would just get all that in there. And that was kind of our, our use case was developing that, seeing if it would work, asking it questions to see if the data

[00:12:49] that comes back is accurate. That was one of the big things was, was testing because obviously you can load all that stuff in, but if it doesn't answer in the right way. And, and that's something that I think that a lot of people don't know is that AI will tell you what, what you, what you want it to, to tell you. You can convince it some things. I feel like it makes everybody a fact checker, right? Like if you're going to use it, the whole idea is it augments your work, but then that pushes

[00:13:16] you to say, well, I'm, I'm the gatekeeper here. And I have to be the one to say, does this even make any sense? Or am I just getting a compilation of data and it's convenient for me to just take it and move on? Yeah. I had a conversation yesterday with a, with a, a senior payroll leader in a mid-market fast-growing organization. And this is one of the areas they really want to lean into

[00:13:41] with Workday is pushing in on the self-explaining pay slips so that they can reduce that first point of contact. And I think most, I mean, maybe not everybody, but most of the generations now expect that you, they want on demand, right? It's not just mobile. It's like, Hey, what else you got? Um, they're, they're not interested in calling someone. So, um, it's just have to, right? I want to ask you a detail about that. I mean, when you, when you started on the agent stuff and just kind of looking at uploading policies and so forth, were you leaning into, um, the capabilities

[00:14:11] of an existing strategic tool that you were using? So I know folks that have done similar things, but you know, they're doing it around capabilities in Zendesk or Workday or SersNOW or whatever it is. Right. So were you leaning into something that was already a part of your technology environment? No, um, I was not involved directly in that. Um, I do know that they were, um, leaning,

[00:14:36] you know, to any of the data that they could pull from, from anywhere, but the, the, the specific, um, areas, I don't know. Um, I haven't gotten into, like you mentioned Workday and I know that they have, you know, some really big AI capabilities that's like outside of what we're kind of talking about. Like, so that's what people have to understand that AI is different. Like there's robotics, there's generative, there's, um, machine, like it's, it's a lot of different things.

[00:15:07] So when you're talking, you have to kind of know which lane you're in. And, um, you know, with, with Workday, I think they're doing a lot of great work where, um, they're analyzing data, like just pulling in a time file, recognizing anomalies. Um, that's a totally separate thing. Um, but you know, when you're developing these agents, um, being able to have as many resources as you can and as many use cases and as many like life experiences, this is what has already happened.

[00:15:35] Like the more data you put into it, the more knowledge, the better it's going to serve. Yeah. Clean data too. Like, I mean, readiness, right? I mean, did you guys have to do any sort of readiness for that to make sure that your data was? Yeah. Yeah. I mean, you've got to make sure that you have clean policies, your SFPs, um, and all of that. That's something I'm not sure everybody gets Julie is, is the part that you need to sort of, there needs to be an AI readiness. I mean, just like we were talking change readiness the other day with Steve. Yeah. Same thing with,

[00:16:03] I think you've got to get your data ready. You've got to get your organization ready, your processes ready. Um, and you can't just turn it over to the bots, right? You've got to have that check and balance that smell test as I call it. So emotional intelligence will always win. Yeah. And that goes back to, you know, this, this beast has been learning from the gazillion prompts that we've been giving it on a day-to-day basis. Um, and so, you know,

[00:16:30] with learning like that, if you want to take it apart and put it into something specific to you, you have to retrain it. Like it knows what it knows, but it has to know your stuff. Yeah. So did you run into when you were working with that and like actively putting out all this, you know, the source and then trying to do that training, you know, one of the lessons we hear a lot is you can't train it by trying to break it. You have to train it by, you know, like pointing it

[00:16:59] in the right direction. So like, how did, how did you deal with that? Like what happened when you were looking at trying to harden it? Yeah, I do that on a daily basis, just when I'm, you know, using it day-to-day, not even, um, in, in like a project like that. Um, but that's something that's really important is that when you get an answer, if you don't reiterate what you were trying to say or point it in the right direction, like you said, then it's just going to say, okay, you liked it.

[00:17:28] I'm going to just do that from now on. And, and that's where all that learning, you, you have to point in the right direction. You can't just say, well, it doesn't know. It's if you're teaching a baby. Yeah. Not everyone's going to prompt us efficiently, right? We all got to learn. So it's a, yeah, trial and error. I love it. I love, I love Julie. I use it to try and, uh, help with our, our transcripts, right. Just summarize our show and help me with our description. Sometimes I organically end up writing all of them,

[00:17:56] but I think it's interesting how it, uh, yeah, I know I'm not using AI the right way. I feel like it's soulless too. I don't really like the way it writes. So I kind of end up writing myself, but I like to see the summary it gives just to see like, did I miss something that I heard or you heard? Uh, so I, I use it in that way, but, but, but you're right. It's like, you've got to, I mean, you can't blame the machine if you're not telling it the right ingredients. Right. So yeah. Pete, when you, when you're doing that, you're talking about just looking at a piece of text and you're not necessarily within the confines of some specific corporate environment,

[00:18:25] right? No, I, yeah. And I'm, I'm literally just using my chat GPT subscription to take, I like to use it personally for me to summarize things. It's just easier for me to take a long set of notes and then maybe say, Hey, I want to get like some points out of that, especially for the show. I'll use it for transcripts. We've even used it over on the source to help us kind of summarize some things. Um, and then I'll write the, uh, write that off of my notes and the, and the thing, but, uh, and their summary, but I think there's, yeah, there, I mean, look,

[00:18:52] I'm not blaming the machine sometimes for not being great at writing it. It's just, am I prompting it? Well, it could be me. And that's, so I'm still learning, I guess is what I'm saying. Yeah. And that's where, that's where you kind of come in and you say, and make, make it sound more human. Like you can say like, dumb this down and explain it to me. Like I'm a five-year-old. Like it's just some of the coolest prompts that people don't know that it's, it's so exciting. It is fun. It is fun. How about when you've got it in your, you know, you're training your baby,

[00:19:18] right. And you have it in this environment. Have you been able to use it to say, you know, how, um, how did my, my policies compare to each other? So am I using the same tone across them or do they all have, you know, six component, you know, like similar components to them or where are their gaps in some of the policies versus our template or something like that? And I think that's a very good question because I think that there are people who will play around

[00:19:47] with it, get stuck and give up because this goes back to giving it the right information. Um, there are not like, so your free subscriptions, not everything will allow you to upload, um, files. Um, if you are in that space where you can upload files, you have to upload the right files. Um, and so there's still a lot to, to come for, you know, uploading a file and saying, Hey,

[00:20:16] can you look at this? Can you do this? Can you do that? We haven't gotten there yet. Um, so it's more so, um, like what you're explaining, um, you know, tell me if I did this right or tell me if this sounds right. Um, or summarize something. Yeah. It hasn't really gotten to the capability where it can actually, um, in that sense, unless you're in the actual application, um, like Outlook

[00:20:42] for instance, or, um, Excel or anything like that, any of the Microsoft 365 products, but, um, yeah, it's, it's not easy. So the only thing I can say with that and then with that question is, you know, if you, if you are out there and you do want to play with it and you do want to upload files and it doesn't give you what you want, take a step back. Um, you know, maybe go on,

[00:21:06] go online. Like I, I swear by YouTube, like, thank goodness for its invention because I can do anything. Yeah. I need to change a tire. Okay. Here you go. Yeah. That's yeah. No, I'm just saying. Like one time I didn't want to take my, my, my two-year-old to the, to the ER. So I learned how to fix, um, I think it's called like nursemaid's elbow elbow. Um, and so like,

[00:21:33] and that's the thing, like go to YouTube, like say, like find a video on how do I, how do I load files into AI and tell what I want? Like use your resources. Don't give up. How much does on the backend of say testing, right. And working with it, does it make you change your processes or change your behaviors that maybe you didn't realize were like, I'm just thinking from a process improvement perspective, is that, is that part of sort of the

[00:21:59] backend outcomes when you're, when you're working with them? I haven't really, um, got like run across that. Um, you're more speeding up processes, making it more efficient that way. Yeah. It was, it was more like, um, trying to get them the, the right answers. Yeah. Yeah. Okay. That's right. You're talking inquiries again. Yeah. Yeah. When you get to the training or I'm thinking like of your instance where you're, you're trying to tell it, no, well, I need to close the loop and kind of tell it what the best answer was. Right. If I'm going to try to train my baby

[00:22:29] and be kind because they may come for us. Yeah. Literally I'm the person, I'm the person that says to AI, could you please? Me too. Me too. I thank my wife's like, why are you thanking her? I'm like, because she doesn't hear you. And I'm like, yes, she does. I don't know. I don't know if it's a she I'm gendering her now, but it sounds like a woman, Amazon Alexa sounds, she'll probably wake up here. So how do you close that loop at the end when you've gotten to a certain point or like,

[00:22:55] let's stick Pete, where he uses it to a certain point and then he's like, okay, yeah, but I like this better. I write that. You then put in the final thing and you say, Hey, for your information, like this is, this is the best, like this would be the best output for the future. Right. Yeah. And, um, I mean, probably doesn't help you Pete if you're using just like public chat and you're not in, well, I have my own license. I don't know if that's private, but yeah, I have my own license. Yeah. If I do write something though, I have, I have tried this. I have said, Hey, tell me if I

[00:23:24] could improve, you know, whatever, improve this and it'll rewrite it. But I've never, I've never come away and be like, Oh, that's so much better. Like I have gotten some words, some phrases to maybe, or rearrange things in a way that made me, gave me ideas, especially with titles. Sometimes when I'm struggling with a title, I'll throw three in there and go, tell me what you think, right? Like just for something to bounce off of. And sometimes it'll shuffle it in a way that I didn't even imagine. And I wouldn't say I've ever come away with a chat GPT title pure, but I've had

[00:23:54] certainly some nudges that helped me, you know, gave me. So it's almost like a springboard for me, like someone to sound off on and say, Hey, what do you think? It's like a test environment for me, really. I haven't gotten to the point of trusting it to entirely do things for me. I mean, you know, analyze this, summarize that like yesterday I have a friend, he's, he's struggling with a, like a, like a class action suit. He was a member of, and he's entitled some benefits and it's this long thing. And I was like, I just don't even, so I put it in and just say, can you tell me about

[00:24:21] this case? Like what, what is it? Why, you know, all that. And it gave me enough to go and look at some of the other things that just speeded up the time immensely. So that's, that I think is great. But then you feel like you're putting that in a safe enough place, Pete, when you're, it's a public case. It's not, Oh, for me personally, I mean, I'm only writing like a, I'm talking about when I write a summary of a podcast, I would never put it in my proprietary client details or anything, but putting in a summary of, of a transcript of our conversation

[00:24:49] and then saying, summarize this for me, what are the highlights? It's more like notes that I then take and use to write, write my own outline or whatever, but, but, or description, but the, but the, but the summarization is happening instantly, you know, by AI taking all that information and just in minutes giving me or seconds, giving me back, you know, five points of what we talked about. And sometimes there are things I miss. I was like, Oh, I didn't, you know, I knew we talked about that, but I didn't think of it in that angle. And it does give me that. So

[00:25:17] in that way, I'm augmenting myself, but I'm not putting any sort of proprietary details. I would, I wouldn't suggest doing that. And I don't know, that's a good question. How private it is. Yeah. That's a really good circle back because do you like, I, I can't preach this enough. Do you not take a file with social security numbers on it and load it into some random strategy, right? Or IP or anything like that. I've heard of CEOs going out there and putting in their, you know, financials

[00:25:44] and saying, tell me this. It's like, ah, you know, yeah. And I wonder though, when you do license chat GPT, do you get a private instance of your chats to where your data is not necessarily going? I don't know. I don't know. Yep. Got to read those terms of service. Yeah. So the company that I'm currently at is they've, they've gotten a license to co-pilot, which is a part of Microsoft. And that's, you know, that little colorful thing. If you're ever

[00:26:09] in Bing, that's the search engine that I've used a lot more recently because it will use AI when you, when you say a search and it'll come back and actually like comb through data and give you a summary without you even asking it. But yeah, so the company that I'm with now when you go to the co-pilot and you know, you're in it because they use SharePoint. And so you can kind of see like

[00:26:33] the box around it, that it's that you're in a secure place. And that's why there's that tab for work and the tab for web. And with the, with the, the homegrown one that we had at my last company, there was a personal tab. So it's, it's kind of one of those things where it's like, you don't want to put your purse, you don't want to ask personal questions and work and you don't want to ask work questions and personal. So you just have to make sure that you know the difference.

[00:27:00] I love that about the Zoho browser. I don't know if you've seen, I can never remember the name of it, but it does, it's got a personal portal and a, and a work portal. So you can be in two different worlds inside the same browser. I'm not sure the others do that. You got to have different profiles. Um, but yeah, that's the same. Yeah, you're right. It's it's, you got, you don't want to mix those things. Yeah. And the other thing too, is that if you're in your work profile, um, and you're in, we'll just say, we'll use copilot for, for a reference. Um, if you're in there and you type in

[00:27:29] stuff and it's been already pointed back to policies and stuff like that, like, so we use PeopleSoft. So I can say, um, I need directions on how to do this in PeopleSoft and it's coming through teams. It's coming through my notes. Um, my to-dos, my, like everything that you have uploaded in Microsoft and it will bring back and it's, and it'll tell you, and then it'll say, you know, I noticed that

[00:27:56] you talk to this coworker a lot. Why don't you go to them and, um, ask some more questions. So it's like, it's from two years ago and now it has gotten so crazy. And so that's why I say like the work in web, because sometimes I'm like, okay, wait, like, I actually want to know, like, you know, about special accounting rule. Like, I don't want to know what, what my company thinks about it. So that's another thing that's kind of happening. So when you're loading policies,

[00:28:24] right. And to let's, let's assume it's, I don't know if it's safe to assume it's the, it's the company being, you know, being co-pilot. Do you have to tell it and help it compartmentalize what that is? Or are you just like, Hey, here's another policy. This one applies to X, Y, Z type of people and situation up. It goes. So the crazy thing about this is that since the Microsoft products are all intertwined, all you have to do is upload it. The integrated. Yeah. Yeah.

[00:28:53] I found Microsoft pilot to be co-pilot to be a little bit. It just, it wasn't there for me. I tried to use it early last year, maybe do some slide creation, some data, you know, data charts. It just wasn't there at that time. You think it's gotten, I think it's gotten better. Yeah. So that's the thing, like wake up tomorrow morning and it'll be better. Yeah. No, agreed. I agree. I did this very early in the year. I think I'll probably come back to it next year and just maybe play with a little bit. But again, I like creating slides my own.

[00:29:21] And so isn't there also a lot. So I was toying with this just last week, maybe, or this week, earlier this week and just something benign, right. That I like to state with really benign things like, okay, I need to pull together some, you know, sales plan for next year. So can you give me a template? You know, can you create a template for a good sales plan? And first of all, it took a dang long time to respond to me. Like I had to keep going back and looking, are you still there? Are you working on this? But then at the end, I just got a,

[00:29:49] you know, kind of poop on you message that said, yeah, I'm not allowed to create any content for you. And that's gotta be a corporate setting. Yeah. That's really interesting. So I have a scrambled brain, like everything is going on in my brain, right? Like I've, there's like 15 chats. I'm thinking about what happened on April 1st of 1992. It's like a hurricane. And so that kind of goes into like, what's so awesome about it for me is that I can say what's

[00:30:19] in my brain, which does not equate on paper and it will put it into something that makes sense. Yeah. And so that's where you start with policy writing, um, disciplinary actions, creating, um, you know, just the everyday stuff that, um, payroll people have to do and, um, that, you know, we have taken so much time, but we don't have time. And so being able to use it for that,

[00:30:45] I mean, that's a life changer. Yeah. Yeah, absolutely. Audits, compliance, everything. So what happens when somebody takes your partially or mostly trained baby away, Amy, you find that you're like starting with a new little fetus or I know it doesn't feel great. Yeah. Well, I mean, the good thing is, is that, you know, going to a new company, um, they, they do have that, uh, integrated with, you know, the Microsoft products and, and all those policies

[00:31:13] and whatever you upload. So, um, I haven't had too much trouble with that. We haven't, I like, there hasn't even been any conversation yet at this company about, you know, building an agent specifically for employees to come to. Um, I'm sure that that might be somewhere down the line, but they're very, very early in their early, early stages. The company I left from co-pilot was

[00:31:36] integrated in, in most everybody's PCs. Um, so now where I'm at, like, and it was co-pilot and there was also their homegrown, like where you could go to the actual website. So there was, it was kind of twofold. Um, but now I just am suckered to co-pilot. Um, but I mean, it helps. I mean, and, and, you know, what we've talked about, it's, it's just one of those things that is,

[00:32:01] it's so new when you think about it, like just as like earlier this year, but it, it learns so fast that it's getting better. Um, but there's a lot of like reprompts. You just have to know, like, Hey, you didn't tell me that right. Like, and you can talk to it like a real person. Well, and I'm wondering, like you were, you were already well in the path to making an agent that could interact with employees, but just like a knowledge base can be a knowledge base. It's

[00:32:30] only used by the internal shared services or whoever the team is and the employees, you know, like, I guess that could be a baby step, right? Where you're doing all of that and you're not yet exposing it to your employees or your managers, but you're trying to do it for your team to help sort through all the noise. Right. Yeah. And that's where it takes, um, you know, a company that's on board and willing to apply the licenses and, and give that opportunity to everybody. Um, a lot of

[00:32:59] companies haven't started or they're just not there yet. They're piloting it. Um, so it's kind of, I feel kind of bad cause I snuck myself in very early. I was like, I can't live without this. Um, but it's like, I can't say to, you know, my coworker, even my director or like, Hey, use it this way because they don't have access. So it's like, there's that too. So there is a licensing bit, right? I mean, like not only does your company have to have their own,

[00:33:28] you know, environment for it that makes it the company or the corporate environment, but they also have to subscribe, ascribe to licenses to the people that they expect will be using it. Do you think your, your company's culture is different in a way that is more innovative leaning than those, some of your peers? Cause I, I certainly see that where it's like, they're very much on the forward edge where some of the companies are just like still on people soft, maybe or not on the cloud, let's say, uh, or some of these other things. Like,

[00:33:56] do you think that that has played a role in having access to it at all? There's a difference between, um, you know, being in a completely corporate environment and, um, and then like the payroll that I've always done where I've always had sites. Right. So you, you have these sites and they don't really like, some people don't even have like cell phone, like smartphones. Um, like are you even use computers? Um, and so there's,

[00:34:26] that's a very fine line between like who needs it and who doesn't. So it's like, yeah. Do you take the energy to build all of this out for someone who isn't even going to use it? So you have to know your audience. That's a very good point. The use cases too, I think have to definitely are going to help dictate that too, because if it's painful for folks and it's going to make it easier, people were going to much likely more lean into that. Right. Yeah. And I, I don't think a lot of people realize how expensive AI is,

[00:34:54] not just to your company, but to the environment, to the, to just like power it draws, right? It's insane. Um, it's, it's really insane. And so that's why you want to be cognizant and you want to be mindful, you know, you could put in prompts just for fun all day. Um, but you know, when you get into like chat GBT four, where it's a little bit more intelligent, like you also have to be mindful and that's why the companies are, you know, still at that stage where they're gauging, you know,

[00:35:24] is this worth the money and the resources and stuff like that? I tweeted a, uh, and I don't remember the source, but you guys can find it out there. I put it on X. Uh, there was a, um, I actually think I got it from the guys at Cohen. It's a chart that shows the, um, the cost savings and the revenue impact. Now I can't remember how they calculate that revenue impact. And I think, I think human resources was like a 40% cost saving 60% revenue impact.

[00:35:51] And they had it for all of the back office, Julie, if you can scroll through that, you'll find it. There's finance, all different parts, but I thought that was interesting, like what it's costing them, but then what they're getting back. And I don't remember how the revenue, I mean, maybe it's per employee or something, I don't know. Uh, or per investment, you know, per project, I don't know, but, but it was interesting to see that they had that. And I thought it was pretty positive for, it seemed for, for, uh, HR. So yeah, but you're right. It is expensive.

[00:36:16] Like you, you think about labor hours and stuff like that. Like you can't even calculate and trying to put it on paper. Um, it takes a lot of work. Yeah. But at some point all of that gets sharpened and these things are running really smoothly. Now you're talking about efficiencies that go flip the other way where now you're saving immense time and immense resources and people are able to do more value added work. Um, that's what I think, you know, the gold in all this is, right. You got to do a little bit of work to get there, right. Chipping through the, through the rock to get to

[00:36:46] the, you know, to the gold, but it's there. Yeah. Yeah. So, so having already been an early pioneer in this, Amy, you know, like where would your heart, I know you might have to re-pioneer some things here as you change from, you know, from one family of baby to another, but where would, like, what really gets you excited? You know, like, where do you think, like, if you just had all the

[00:37:10] time and you could pull these things out of your, your brain and grab a couple that you're just itching to think about, where would you go with it? Yeah. Um, you know, the thing that most piques my interest is the involvement of integrating into our payroll systems. Um, I think about exception reporting and the amount of time that we spend, like not everybody has caught a product like Cognos

[00:37:35] where you can bundle a bunch of reports into one, um, you know, where we're at now, we have to, we can batch our queries that we run, but it's, you know, each individual one. And so, um, just think about a world where, you know, you have these anomalies, like, I mean, payroll errors, that's like something that we never want, right? Like we never want it to be our fault and we want to catch the

[00:37:59] things, but then we don't know the things and like, imagine a world where, um, you know, you've used it enough or you've trained it enough where, um, where it knows what to expect or what isn't right. And, and things like that. I mean, that, that's like my, my big brain. What would be so cool? It feels like that would be a major shift, right? Like when you're talking about policies and agents,

[00:38:25] you're talking about a lot of word language, right? And it feels like, um, yeah, yeah. It feels like tackling exception, you know, and identification becomes much more spreadsheet numerical types of stuff. Right. Um, and, and who knows how, you know, it would treat that. And if, if there's AI, you know, ways to do that, there certainly are, um, versus robotic process automation,

[00:38:51] what you're talking about, but, um, but different tools for different purposes. And I, I think, um, I wonder all the time, just even when I was fiddling around with the sales stuff, like, what if I tell it, like, here's, you know, like, here's the grid of where, you know, how I think I'm going to break this down and achieve this. Can I, if I put that in there, will it know anything smart about it or not? Right. I mean, or is it just to really know what to do with it? You know what I think is going to be really, really exciting for HR, especially, and

[00:39:20] absolutely for payroll, right. Is like, is the, is agentic AI, right? The ability for agents and co-pilots and well, co-pilots really assisting you and agents actually going and doing the work for you. If you've got multiple agents out there, right. Think about it, right. Any process payroll has got a hundred checklists. HR has got a checklist. Recruiting has got a checklist, right. On and on and on down the line operations and whatever. And a lot of those things overlap and touch each other. And some of those things are in the same system. Sometimes you're all on the same

[00:39:47] ERP. Most of the time you're not, especially as you get a bigger company. What I'm excited for is seeing those agents start to talk to other agents and doing that checklist to checklist handoff. That's making sure these processes, like we talk about and beat up leave, leave of absence all the time, all the handoffs. But there's many of those, right. Benefits. You think about onboarding, right. Just the onboarding experience. There's someone in HR involved from recruiting all the way

[00:40:12] through in the onboarding experience, everybody's sort of being touched. Right. So all of that needs to be woven together. And I think that's the, the, the agentic AI is really what's exciting because now you're talking about eliminating those checklists, stringing together processes and really, really changing the, the, the value proposition in that, in that experience. Having time to get up and take a lunch. Yes. Right. You can take, yeah. Payroll can take a Christmas off. I, everybody's exhausted. Everybody I'm talking to right now on payroll.

[00:40:40] I've never forget this, but I've, I've lived it is everyone's exhausted right now because it's, you know, and we still are just, we're, we're just starting the beginning of really year end. If you think about what we still have to go. You can innovate your way to an anniversary dinner. Yes. She was telling me before that, before she really knew what it meant to work in peril that, you know, like December 30th is not the best time to try to celebrate anything.

[00:41:06] Man, I've missed a lot of things. Yeah. Yeah. I got married at year end. I don't know. It's what you sign up for. I think in that, in this game, right. You gotta, you're, you're the people who got to keep the lights on, right. Someone's gotta, someone's gotta do it. So yeah. Yeah. And then, you know, in, in the spirit of your end, like how cool would it be to have a system that was so intuitive that you go to enter a negative deduction on a, an adjustment or a check or something. And it, a box pops up and says, Hey,

[00:41:34] wait, if you do this, your W2 won't produce or something like that. Like, I mean, it just, it seems so tiny, but. Or, or how about the, the ability? I mean, if you think about it, really, you're almost having to conduct year end every quarter, right. If you could at some scale conducted in a way that's the, that has the AI again, again, I think very agentic going across all these things and you're right. It's nudging you and surfacing things to you in the quarter or in the month and every moment to say, fix this. So you don't have that big mess at the end that

[00:42:04] we're all ticking and tying. So. Yeah. So it's not just taking our jobs. It's, it's, I mean, I look at it as making our jobs better. Yeah. Yeah. Yeah. So, so those things make me just think, well, should it be coming from, you know, initiatives or centers of interest? Like you're creating Amy that are, that are trying to train and create something that will think about this, or is this really what the payroll provider should be creating and embedding inside of their tools?

[00:42:32] If they even can, right. If they, they're subject to their own work environment or their own, you know, universe of. They are sure playing around with it. Small language. Yeah. Yeah. That's where your analytics come from. Yeah, I know. Right. Like, but you know, like, will they get there fast enough? Will they, you know, have enough data in their own walls to be able to do a good job of that? And, and it might actually impact how you, um, how you put together your

[00:42:59] service delivery model and what things are important to integrate and not integrate, you know, and, and keep within the same system and what things are not because, um, you know, there are limitations if you're, if you're bouncing out to other, other universes. Right. Yeah. I think everybody in the payroll world, especially in the systems world is just sitting back waiting. Yeah. Yeah. Who's going to get there first. Oh my God. That's a good point. Actually. Would you

[00:43:25] have any advice to anyone? I feel like there's, this is as much a bottoms up sort of thing than it, than it, as it is a top down, meaning like, don't wait for them to just drop AI on you or go get AI. Maybe there's a bottom up piece to this, like any advice to, to other payroll leaders out there that kind of want to nudge their organizations to be a little bit more innovative and engage these things. Yeah. I would say start, you know, start with your technology officer or somebody in that,

[00:43:50] in that space, because it could be applications. It could be product. It could be, you know, maybe somebody on the help desk that's just interested in getting it started, but you definitely want to start in that technology space and, you know, just say, Hey, like, I think that this would be really, really awesome for us. And the thing about executives is that you really have to build that business case and, you know, where you were, where you were talking about

[00:44:18] the article on X and saying, you know, this is the cost savings and the revenue and stuff like that. I mean, it's finding those, those things that you can reference back to and that you can really, I mean, you really have to plead your case at that point. Yeah. Yeah. Yeah. We've talked a lot about, about using your data and making cases. And I think you're right. If you can show, look, this action would move the, you know, this investment in this use

[00:44:44] case would move the needle by this. It's, it's changes the game, right? People start looking at that, especially if the dollars are talking about it. Yeah. People are motivated by money. Show them all their save money. Especially C-suite. Yeah. And from a cost center, right? I mean, payroll is a cost that if you, you know, at the end of the day, it should be returning value, but, but fundamentally they want to save costs. I mean, that this is how you do it, right? Start knocking down that leakage time, whatever, you know? Yeah. And that makes me think, you know, like, what if this, this world that we're in now and this AI and

[00:45:13] these capabilities, what if we could prove, you know, that we, that we do have value, you know, I think that's been something that we've been talking about a lot lately in the payroll world is finding that seat at the table. Like what if at the end of this, we aren't a cost anymore? Like, obviously we do create a cost, but, but us as people aren't a cost. Yeah. Yeah. No, you should be in an agility enabler, which is the, the ultimate, you know,

[00:45:41] ROI that, that you can give back to the company. And companies that aren't seeing that and investing in payroll are really, I think at a competitive disadvantage. I actually think I proved that with the, uh, uh, the payroll profession confidence index to some extent that showed just the impact, the max impact. It was really, really visible. Yeah. Yeah. So what, what if you want to talk about what else, what if we met payroll with the criticality, like it met the criticality and the

[00:46:06] importance with the innovation and resources? What, what could be, I mean, you guys do an amazing job with, you know, one hand tied behind your back. What if you actually had help? What if you, you know, had some air Jordans and you could, you know, sprint, you know, come on. And we just started like 10 projects in our mind in that very moment that you said that, Pete, I could tell. I know. I'm very inspiring. I think about how, you know, we're always being pulled to the next fire and we never really have time to innovate. And so it's just that innovation that's been sitting in the back of our minds

[00:46:36] that we just, we know make it better. We don't have the time. Yeah. Yeah. Yeah. Well, yeah. Right. You got to fly the plane and, and fix it and make it, make it bigger engine on it. So yeah, no, I understand. It's like always band-aids, band-aid, band-aid, band-aid. Right. What if we were able to actually dissect it, take it apart, make it work right? Yeah. Agreed. What if? Awesome. Awesome. What if? Oh, well, I know we can think of a lot of things from here on and I'm sure you have so many more

[00:47:06] learnings, Amy, and thank you. I feel like we better bundle up this particular, this particular podcast episode and then see when we can check back in with you and just talk about whatever new things are hopping out of your brain and into the, you know, into the world of payroll operations and see where you're at with it. Yeah. Where can everyone connect with you, Amy? Any, anywhere links you'd like us to share? Yeah. You, I mean, if you just search my name on LinkedIn, that's where I'm most

[00:47:34] active. I never signed onto the Twitter. I think it was past my time and not part of my, my needs. I'm with you. I'm not great, but then you can just search my name on LinkedIn. You'll find me. Yeah. We'll share your link as well. So folks can find you, but yeah, thank you so much for coming on. This is great. It's always good to talk to practitioners, listen to your mind, you know, what, what your lens is here. What's going on. It's exciting. I'm so excited for payroll right now.

[00:48:00] All the stuff that's out there, man, couldn't it's long overdue, long overdue. So that's good. Okay. Well, look, we are on a wrap this one up and look, thanks to you again, Amy and Julie. Thanks as always. And thanks to everyone. We'll we'll see you all soon. Sounds great.