Steve 0:09 The HRHappyHour Network is sponsored by Workhuman. Employees recognized for milestones, are three times more likely to believe their company actually cares about them, three times. So why are companies so bad at it? The stale bagels, the branded swag, the audacity. Workhuman believes milestones deserve their moment, the space for the people who know you best, your actual work circle to reflect on your journey, your wins, your impact. The inside jokes, the big moments, all of it. It's called service milestones. AI finds the right people. Automation handles the rest and employees choose their own reward from millions of options curated just for them. Milestone like you mean it with Workhuman, a proud supporter of the HRHappyHour network. Find out more at workhuman.com and thanks for joining us. Welcome back to the At Work in America show. My name is Steve Boese. I'm with Trish Steed. Trish, what is happening here? What's going on? Trish 1:06 You know what? I am in a good mood because I now have two children who have both graduated college. So congratulations to my son, Jack, and co host of The Play by Play podcast. But he graduated from University of Utah this past weekend with his degree in economics. Steve 1:24 Thats pretty cool. Trish 1:25 And I don't want to spill the beans. I'm gonna let him spill where he's working, but I will just, I'm gonna tease it. Steve 1:32 Don't jump in front of his announcement, Trish, please. Trish 1:34 He has a full time job, though, with someone that is a friend of the show. Steve 1:39 Okay. Trish 1:40 So, yeah, more to come on that. But no, it was a good weekend. How about you? Steve 1:43 Yeah, real good. Really excited to be here today. It's a topic that we've been talking about a lot, in general, right? Which is, of course, AI, right? It's on every subject, every presentation, every conference, every probably HR teams, weekly meeting planner and, of course, all over the HR tech industry. But we haven't talked enough about how it's actually showing up at work, how it's changing, how people are getting work done, how it's improving processes, driving efficiencies and actually enabling people to focus on sometimes higher value added work. So we're going to talk about all those things and maybe some others. Our guest today is Tony Truong. Tony is the VP of people, strategy, operations, technology and analytics at Chime. Tony, welcome. How are you? Tony Truong 1:43 Doing great, Steve, thanks for having me. Hey Trish, how are you? Trish 1:43 Good. How are you? Steve 1:43 Chime is a super interesting company, and so first, before we dive into talking about AI, how it's showing up at work, some of the things you guys are experiencing with AI, maybe tell us a little bit more about you, Tony, and then give us, you know, 60-90, seconds on Chime. Folks probably know Chime, but maybe not know everything chimes gets involved in. So we'd love to learn more about that too. Tony Truong 2:04 Yeah. So I'm Tony, been with Chime for about two years, and like you mentioned, I lead the strategy, ops, technology analytics team. We call ourselves the sota squad, S, O, T, A, and we are here to develop the system to drive people process technology and data transformation of our people team, and also, what does that look like you know, for the company? Chime is a great company. We are as you know, our company focus is to help everyday folks unlock financial progress. So we do checking, savings, credit building, early wage access, really serving the everyday Americans. The two thirds are living paycheck to paycheck. So very mission driven, and I'm proud to be here to support on that mission. Trish 3:51 Yeah, it's such a great mission. And I know, you know, Steve and I have had many briefings over the years. It's, it's one of those where we're proud to have you on not just because of the topic. But yes, Chime and what you all represent and really are, are helping a majority right of people in this country that are living paycheck to paycheck, who, in the past, didn't have a solution to help them, sort of get access. You mentioned, like early access, things like that, are really impactful. So just congratulations on Chime and all you're doing to, you know, to really make, make the money you earn more accessible to your employees. Tony Truong 4:26 Absolutely. Steve 4:27 Yeah. Chime's an inspiring company. It's won a couple of awards. I've been a part of doing top products of the year awards, and just doing really cool things for people and for organizations as well. So what we want to talk about Tony, though we've been just every day I talk about AI. I see demos on AI. We read about AI, not just me, right? Everybody, but I'd love to learn a little bit more about how you're unlocking AI, actually, practically, and having the seeing impact from AI in just normal kind of people operations in HR workflows. I'd love for you to maybe tell us a little bit more about that and what you guys are seeing at Chime. Tony Truong 5:06 Yeah, absolutely. AI is really changing everything that we're doing, and it starts with just fundamentally what does it do for us? And we look at it as removing friction, you know, removing all the high volume, repetitive things, and making it more systematic, automatic, and having being a co pilot assistant to every person here that's using the tool. Two specific areas that we've targeted, at least on the people on the people team here is one in recruiting. You know, in the original state of recruiting, a lot of things were highly manual, writing job descriptions, creating interview plans, filling out the scorecards, right and also the interview itself, like you have to take notes. You have to translate that into, you know, documents on how well the candidate did, and then, like, take that into scorecards, and then have a debrief about those candidates, and then make the decision. And a lot of those things aren't even documented. It's all on someone's desktop. You know, we don't have any technologies that's actually transcribing, recording or making sense of the data. So one of the first things we did with AI was actually introduced a recruiting co pilot, okay, where the recruiting team works with every hiring manager on every position at Chime, understands what they're looking for, and AI actually helps generate the first version of the job description. Of course, we do some of the human touches on it aligned to our values, what we look for specifically, but now it actually drives a majority of the content development around job descriptions. And then from there, it can understand who needs to be part of the interviews. So the hiring manager, the recruiter, works on the slate of folks that need to be on, and they develop interview questions based on competency, values and things that we care about at Chime that makes a great candidate here, that then gets pushed into the interview process. And now we have an AI agent that actually pops up in every interview and does the recordings of the interview. So most of the time, it's hard to pay attention in interviews when you're taking notes, asking questions, and now the interviewer can be hands free and just really focus on a candidate and the questions. Steve 7:37 And it's hard sometimes to look back right when you're maybe doing an internal circle back with the team, right? To say, oh, to remember everything that was maybe said in the interview or anything of note, right? Especially if you're doing a bunch of interviews and you're doing your normal job too, as a hiring manager or participant on a hiring team, having that AI agent in there to record and document and make sure nothing gets missed, I think that's got to be a huge benefit? Tony Truong 8:01 Yeah, absolutely. I mean, it takes a lot of the things again you do that are mundane, repetitive, and actually just automates a lot of those things for you, and then it pushes it into the system, so we have it automated into our our ATS or applicant tracking system to translate all the recordings into notes about each candidate, and then the scorecard is automated, and the person just needs to go in say, yes, that's reflective. Thumbs up, thumbs down. And that's been the biggest drag. Was actually that whole process from creating the job descriptions to the interview plan the scorecard, so we're able to reduce a lot of that from weeks, maybe even sometimes a month or two, to now, like hours or even days. Wow, so the productivity has been significant. Trish 8:55 Yeah, I made so many notes as you were talking because, again, having worked in HR and Steve probably might agree. It's like there are so many little things that you mentioned that are huge, huge impacts. So when you talked about having the data available, not just on someone's desktop, I can't even tell you in my career how many times leaders have left the company, and we've lost all of that information because it wasn't in a standardized place in the ATS. It wasn't captured that way. So that alone, I think help is so helpful. Also, just when you said, you know you're you're basically given something to react to. That scorecard comes to you, and then you're sort of adjusting, making sure it's okay. I think especially once you get to be manager and above, it's much easier to react to something instead of create something from nothing, right? So not only a time savings, but also just an accuracy check, I think too, that we wouldn't have had in the past, right? I know that it was always like pulling teeth to get HR leaders. I mean, HR managers. um, to work with the hiring managers and get all of that feedback in a in a really impactful way. So I think just those two examples alone make it well worth having an agent sitting sort of beside you right. What kind of feedback are you getting from the people actually using the agent now that you have that in place? Are they open to this? Do they find it helpful, or do they still kind of hold on to the you know, we know what we're doing, even if it is mundane. Tony Truong 10:27 Yeah, I think initially, there was not so much resistance, but just trying to understand how this is going to be impactful and helpful. Because before they had actual recruiters that were providing white glove service that were they were interacting with someone to actually go through the process with them. And in the early days of AI and large language models weren't completely accurate. So it took time for us to change the it's not so much about implementing the tools, about changing the operating model and the process. So once we are able to, you know, define what that was, drive governance towards it. So it wasn't an option to go white glove service anymore, like you have to go through this new process, and we're going to make it better and refine it. It actually was a huge productivity lift for everybody, and also unlocked a lot of asynchronous collaboration, because now things are recorded, things are documented. You don't have to set up a meeting or conversation with a person. I remember, I was interviewing a candidate, where actually my team members interviewing candidate. I thought, look good on paper. She interviewed that person. And was like, oh, this person was okay. And I was like, hmm, that's strange. And they sent me, actually the video, like, go watch the video, go review the notes, and I did it. I just need to set up a meeting. I got the content directly, and I was like, Okay, I guess you're right. This person isn't really the right person, so that actually saved me a lot of time, as you can imagine, with setting up a meeting, having a debrief with that person, understanding meticulous notes, but all that was documented, and AI was instrumental with that. Trish 12:07 Yeah, I can see that would be a huge help, too. I mean, I think even after sometimes people come on board and then they're like, this wasn't what I agreed to, right? Or something, you could go back and look and see exactly what they said and what was told, you know, during the interview process. So yeah, I think many uses for that type of recording, I would imagine. Tony Truong 12:24 And the analytics behind that super powerful, like, you can look at the ratio of interviewer to interviewee, you know, talk time to see our. Steve 12:34 That's a good one. I always like looking at that one. Tony Truong 12:37 Yeah, you can see how often people are mentioning compensation culture, like, what really matters to the candidate, even the compensation package of stuff like, what are we offering? What are their expectations? All that's being documented so it helps us make better decisions around our compensation practices. So the wealth of data just from all that is super helpful, because at the end of day, the data really helps drive decision making, and if you don't have the data, it makes everything else harder. Trish 13:11 I feel like that would have been so valuable to have years ago, because going to your CEO or your other C suite members, and all I could really talk about was productivity, right or time to hire, or something like that. You've just listed three or four things where that would have really made HR seem more valuable in those discussions with the, you know, the C suite, I think too. Tony Truong 13:32 Absolutely. Trish 13:33 Yeah, really. Steve 13:33 Tony, are you seeing like, yeah, Trish mentioned time to hire. Look, there's always, there's, you know, a whole bunch of standard recruiting we're testing. We're talking about recruiting, recruiting recruiting metrics that folks track and have been tracking for a long time. Is there one, are there ones you guys look at? Or did look out specifically to say, hey, if we change our workflows and incorporate AI into various steps of the process, we're going to we're going to look for these outcomes. Is there any analysis like that you're doing, and if so, like, what are you finding? Tony Truong 14:03 Yeah, we look at the recruiting cycle time. So everything we do with AI, we reduce that to about 20% faster. Steve 14:11 Wow. Tony Truong 14:11 Again, like taking things that take days weeks now until like hours, right? Or even, like minutes, we look at time savings per role. We saved about 10 hours per role, which is huge especially while you're moving quickly as a fast growing company. We look at cost per hire like at an end day, like, what are the inputs and outputs when it comes to hiring that matters a lot to us in recruiting. And we also look at time to fill. Are we moving things or we're moving roles quick? You know, are we able to find the right candidates in a timely manner, as when we need them? And that's part of our broader workforce planning metric, as well as, like, finding the right talent, right place, right time? Steve 14:55 Yeah, that's great. Thank you for sharing that. Tony. Trish 14:59 Yeah, I think also, like, when you're talking about this in terms of accuracy, you sort of mentioned, like there was some trepidation, maybe, you know, a few years ago, with accuracy of these inputs that are coming in, how have you seen that change now that everyone is using it, and you're actually getting some real results filtered back to you after the person's on board. Is the accuracy where you need it to be? Or is that something that's still a little bit of a work in progress? Tony Truong 15:27 Yeah, I think anything has to do with data and AI is always going to be work in progress. But the models itself have gotten so much better, the AI models and the data has become more and more refined and richer. So it's getting very close, if not almost very accurate for us, at least, of how we're using the data. Trish 15:49 Yeah, that tends to be a thing that HR leaders, especially are fearful of, right? They're sort of putting their, you know, their reputation, on the line for saying, we need to do this, and then if there's inaccuracy, so that's good to hear that you're actually seeing data to support the movement in a positive direction there. Tony Truong 16:07 Yeah, when it comes to AI, it's not just working with AI or having it work for you. It's actually you got to work for it as well. Trish 16:16 I like that. Tony Truong 16:17 Yeah. Meaning, yeah, it's a partnership. But also a little bit of like, you know, servant partnership. If you want the AI to be smarter, you got to feed it with great data. You got to refine it so that it can become smarter for you. So that's the feedback. Steve 16:33 Yeah, that's a really good point Tony, and I think, an underrated one about AI in the enterprise, right, where we've got to spend time and work as an organization, as individuals, even too to help the AI get better. Help it. Help help it help us, if you will, right, drive better outcomes. You know, I was at an event last week, yeah, last week. And, you know, one of the big HR tech providers, and we've been hearing so much about agents right across the industry, across HR tech and enterprise tech in general. In fact, this particular vendor said we're no longer going to even talk about how many agents we have. It's ridiculous, because there's there's so many now there's like hundreds, right? And it's become hard to even keep track of and I know we talked about recruiting some there are agents across the board now available right for HR organizations, for people, ops organizations. Are you finding other areas across the employee life cycle where, hey, we've been able to introduce AI or agentic AI into our process, make some modifications and see some really interesting outcomes at Chime? Tony Truong 17:43 Yeah, our the second use case, I would say, has been really instrumental for us, is around our service delivery. Okay? We're providing support to every employee at Chime, you know. And prior to AI being introduced, it was Slack channels. It was reaching out, emailing your HR business partner. It was, you know, a swarming method. Anyone that was in the Slack channel would respond to it, yeah. So we had to introduce a tiered service delivery model approach that was more AI first. So starting with Tier zero, like building a very strong knowledge base, so that when folks have questions that they need to ask, that things will get served up to them in an automated fashion. They can self service themselves. And at Chime that it was very, it was a successful implementation. 75% of all of our inquiries were resolved at Tier zero and at tier one. Steve 18:49 Essentially with the AI, right, Tony, not, not a person directly involved then, Right? Tony Truong 18:54 Correct. Okay, correct. And then a tier one and twos are, you know, folks that it's essentially our share service model that needs. You actually need human judgment and coordination involved, like recruiting coordination, system transactions, onboarding, offboarding, or you need to talk to a specialist about your benefits, you know, or things that are strategic around org design, workforce planning. So we really build a, you know, a service matrix so people know the channels and paths to go to versus, you know, the very disparate paths that we had originally with Slack channels, emails or just reaching out to someone. Trish 19:34 I think it solves some of the problems of self service as an idea in general, because if you go back 10 years or so, people were really resistant to any kind of self service. You mentioned white glove earlier, right? And I've heard that a million times at different employers I've been talking with, and I think that when you're talking about that, so many of the resolutions can come at that tier zero, that in and of itself is white glove service to those employees. You just can get your answer right away, right? That's so much easier than having to call and go, you know? It just takes so much longer. And then it also, I would imagine, boosts the care and the white glove service you're offering to those people that fall into tier one or tier two, right? So it feels like it would be a win. Win. Are you seeing that from the employees who are, who are going through it sounds like a, definitely a win for HR, right? But what about the people who are actually, you know, the 70% that are solving it at Tier zero? Are they feeling cared for? Tony Truong 20:32 Yeah, absolutely. And I'm glad you actually mentioned that, because we actually reframe white glove to being more more AI focus, because it actually does make help answer things faster for you, right? Trish 20:46 Yes. Tony Truong 20:47 You get what you need quicker. So at the end of the day, we want all the operational overhead that people's team, you know, services team, does for employees, to be as smooth and as frictionless as possible so they can go back to doing what matters most, which is like certain. I'll say some of the headwinds we had in beginning, you know, were around the change management because they were so used to working with somebody. Steve 21:15 I'm glad you mentioned that Tony, my notes I wrote before we started talking, I wrote change management on my sheet here that I wanted to bring up and ask you about. So I love that you brought it up before I had a chance to ask you. Tony Truong 21:27 Yeah, change management. I think one again, it's we always start with governance, just making sure that the workflow changes or the operating model change matters the most, that we actually have leadership support and we actually role model this, you know, as a as a company. Second is around your manager readiness and leadership readiness and making sure managers and leaders are all trained, and we work through our HR business partners to drive that. Third is like coms, making sure that we have clear messaging and support on how do we land the what, why, when and how changes, and that actually pushes through across every parts of our organization. And I think what's most important is not just when we launch, you know the solution or change, but it's actually what we do afterwards. And we will, we will we do with the data. How do we work on continuous improvement of the process and iterate and make sure that we are evolving and improving? You know, the work, so that matters a lot to us, as much as the launch itself. Trish 22:34 You mentioned the internal coms. Do you and the HR team have input then on how AI is responding like so that, for example, I'm thinking, you know, couple employers, we have your own acronyms. You have your own sort of language of business within your organization. Do you all have input on how the AI responds in terms of making that match the cultural language that you have at Chime? Tony Truong 23:04 Yeah, so our knowledge base we we make, we call it like making things Chime ified, speaking our language aligned to our values and really serving the outputs that are, are things that we care about and what we don't want it to do is, like, hallucinate and say its own I don't know its own findings or its own logic, but we actually point it to content that it should be referencing, as you know, the single source of truth for for our content? Trish 23:38 Yeah, I think that's so important, because especially the longer you're with an employer, you really do sort of buy into that cultural language, right? And so if, if AI was giving you something different, or even just more general, it wouldn't feel like a partner, I wouldn't imagine, right? You sort of think of like an agent is sort of like your co worker, right? And that they would be speaking that same unique language. So I'm glad to hear that you all have really, like been able to input on that as well. Tony Truong 24:08 Yeah, and we look at the data. We look at are things being escalated outside of tier zero? You know that they need to talk to someone. We look at adoption and deflection rates. Are people going through our AI bot to answer, to get their, you know, their questions answered, or are they still reaching out to people and again, we we look at continuous improvement, like, how can we make sure our knowledge base is refined and and up to date as much as possible so that it delivers the speed of content that people need. Trish 24:45 Yeah, without getting too, too personal. Is there any group like as you're seeing, like, who's who's still sort of reaching out to a human right, versus using an AI first? Is it a certain type of role? Is it a certain type of generation that's still finding that, or is it just kind of a mixture of people who, for personal reasons, might still go out and reach out to a person directly? Tony Truong 25:12 Yeah, I would say first and foremost would probably be someone that is very early to Chime like newly tenured, and they are still figuring out where to go for things. And I think as humans, we just tend to want to just email or reach out to someone. Trish 25:29 Yeah, right. Tony Truong 25:31 But we actually do, as part of like onboarding for every employee, introduce all the different elements of support, and the first line of defense is like that tier zero agent, which we call Optimus Chime here. Trish 25:45 Oh, I love that. Oh my gosh. Tony Truong 25:48 And Optimus Chime is not just plugged into the people, it is plugged into other parts of the organization. So it's truly the one stop shop for anyone that has questions. And it's intelligent, smart enough to know where to route or to reference for information. Trish 26:05 Steve, I mean, I can see, like, if we had something like that back when we were in, you know, really in the trenches of HR, it was like you'd point people maybe to your intranet, which, the search never worked, right? It never brought you what you need to. I can't imagine how much time we'd save. Steve 26:21 Yeah well, right? Any, one of the major challenges with any kind of knowledge base or repository was getting people to use i, one, keeping it current, two, making it usable, three, there's probably a couple more I can think of, but those intra nets were fraught with all of those challenges, right? Modern approaches to service delivery and serving up knowledge and information. You know, counter just about all those. I think. Trish 26:47 Yeah, I want an Optimus Chime. I want it to be easy. Steve 26:51 Tony, I have one last question. I think for you, you really eloquently described some of the keys to adoption, right and getting people on board and talking about communication and feedback, et cetera. I'd love for you, if you have any just thoughts, advice for we have a lot of HR people listening to this show who might be wading into these kinds of projects in their organization or considering them. Do you have any one or two things you guys have learned along the way that for an HR person or an HR organization, HR leader, hey, this is good to keep in mind, good to think about as you start to incorporate AI and maybe transform some of your workflows? Tony Truong 27:31 Yeah, I would say you really think about the operating model of the HR function as a whole. Is it siloed? Do you have the skills and the subject matter expertise you need? And think about the org design and operating model first, because that's really the foundation, what I'll call like the system of or how you want the team to operate as a system, as an HR organization. And then get into the workflows itself, like, how is work getting done? And the two things we always ask is, are we just refining and sort of making incremental improvements on what is happening today, or are we actually transforming and rethinking how work is actually getting done with technology and AI in mind, and we don't try to chase the shiny object, like, let's build agents to replace everything, but we actually look pragmatically at, hey, does a workflow where from candidate to hire? It requires a lot of manual steps. Can we remove a lot of these manual steps and have synthetic or AI capabilities to replace certain steps, and where do humans still need to be involved? Steve 28:46 Yeah. Trish 28:48 Okay. Tony Truong 28:49 So those are things I would recommend really think about the operating model, the org design, the org structure, you know, how teams should be working together, and then think about the workflow itself and reimagine the work versus just incremental improvements on refining what works today. Trish 29:06 My last question Tony is, when you're doing all of that, are you including leaders or even employees from other parts of the organization, or are you doing this with just all across HR, how are you? Tony Truong 29:24 All of the above. Trish 29:25 All of the above. Okay, good. Tony Truong 29:26 We definitely you. We always say the answer is, always with our customers. So we always think about and we develop the personas of who are we servicing and what are we optimizing for. So we definitely do have a test group that we always pilot with, and we always like to dog food our own process within our own so we do that as well. Trish 29:50 Yeah, thank you for sharing that, because I think that's one of the things, if I'm, you know, listening to this show, I'm thinking like, Okay, do I do it, you know, just with my team, or do I go across? So, really good support there. Steve 30:01 Yeah, Tony, this was really fun. I appreciate you sharing about your experiences there and what the teams have been doing at Chime. We're turning AI into a really actionable, practical and impactful tool, right? It's not it doesn't have to be a mystery, it doesn't have to be scary. It can it can. It can be done. It can be done effectively and can have real impact. And I appreciate you sharing some of the experiences there at Chime. Tony Truong 30:25 Thank you, Steve, thank you. Trish, it was fun having this conversation with you guys. Steve 30:29 Yeah, we appreciate it. So we'll encourage folks to check out what Chime is doing. Chime is a great company doing really cool things. We touched upon some of those things at the top of the show, innovative products. Really great mission, helping lots of people. So we're fans, and thank you Tony and the team at Chime as well. So Trish, good stuff. This is super fun. I dig AI, but I like we talked about AI in a way that was understandable and approachable, and hopefully folks will take that away from this conversation. Trish 30:57 Yeah, I agree. I think, you know, whenever I'm listening to any podcast, my I always want some takeaways, something I can go back and do differently in my organization. And I feel like, Tony, you've really brought so many just different ways to think about things and reframe how people are handling it. So this is really helpful. Thank you. Steve 31:16 All right, great stuff. So we will leave it there for now for our guest, Tony Truong. Trish Steed, thank you. My name is Steve Boese. Man, great show today. Go check out hrhappyhour.net for all the show archives, tons of content. I can't even get into it all. Look for this special top secret Jack McFarlane announcement on The Play by Play coming soon, and thanks for listening. We will see everybody the next time, and bye for now. Transcribed by https://otter.ai