Empowering Supervisors: A New Lens for Workforce Engagement

Good morning, everyone. This is Vicky Harrell, executive director of SWPP, and welcome to our webinar today sponsored by IntraDiem. We’re gonna teach you about how to empower your supervisors and give you a new lens for workforce engagement. We’re excited to have you with us today. And, we’re going to record this, and we’ll send out the recording after the session. So if you wanna share it with anybody else, you can share it with anybody who wasn’t able to attend today. I also want you to interact with the panelists today by using the chat on the bottom right hand side of your screen. And if you would look at that chat box, make sure it says to everyone so that everybody can see your comments as you put them in. We do wanna start with one question just to get you started in the chat. If you would put in there, maybe where you’re listening from, but also how big is your contact center and how many agents are you supporting at this point? So let’s put that information in the chat and get us started in the chat. Right. And I hope everybody is having beautiful weather. I’m certain there are some people up north that are not today. I have been complaining about the weather in Nashville forever, and now we’ve got gorgeous weather coming this week. We’ve got seventy degrees tomorrow, which I am excited about. So We’ve got the sun up in here for the first time in a few weeks too, Vicky. It’s nice. Oh, good. Good. It, it has been it has been gray and dreary and cold here, and we had snow last week, which was awful for Nashville because we don’t know what to do with snow. So That’s right. Alright. It looks like everybody is finding the chat. Again, that’s where how we want you to interact with the panelists when they ask you questions and want some responses. Or if you have a question as they go along, make sure that you put that information in the chat. I’m gonna turn it over to our two presenters today. We have, Josh Wilkins and Tristan Pahud from InterDAM. I’m a turn it over to you now, guys, and let you get started. Welcome. Awesome. Thanks so much, Vicky. Hi, everyone. Nice to meet you. I’m Josh Wilkins. I’m a solutions consultant. My job is to kinda bring the product to life, explain how the rubber meets the road, what to do in the last mile, and the the the details on launch and all that kind of fun stuff. I’m joined by Tristan Paoos, our product manager. Tristan, please introduce yourself. Hey, Josh. Hello, everyone. My name is Tristan Baird. I’m, from France and based in London in the United Kingdom. And, I work with Intradiem as a product manager responsible for all AI and machine learning initiatives. I’ve been with the company for, three years now, I believe. Yeah. Awesome. Thank you, sir. I’ve been here almost four years now, giving the timing. That’s nice. I’m a former customer myself. I used to use to manage Rogers Communications, the, Intraday team there, so I’m familiar with using the product. So any kind of really specific oh, I’m saying I’m not hearing us. Audio check, you guys can hear me okay? I heard Tristan okay. Yes. You sound good. Okay. Perfect. I’m gonna leave the chat to Larry then. Thank you so much. Okay. Perfect. So what we’re here talking about today is, contact center burnout or or, the the cost of attrition. So, and I’m gonna do my best to speak slowly. We’d like to kind of get to the details of things. We get a little bit excited, so please keep me honest if I jump past that. So speaking to our customers, we really understood that we’re sitting on a mountain of data. Intradiem is an intraday day automation platform. We sit between the ACD and the WFM systems, kinda reading information in real time, understanding what’s going on within queues and at agents’ desktops, and then doing something about it, taking action, delivering an extra break, delivering training, seeing if somebody’s stuck in ACW, for example, and giving them some sort of outreach to solve the problem. Because of that integration, because of those integrations, we’re sitting on a mountain of data. And speaking to our customers, the idea of an employee battery was born. Because we process all of these transactions annually, we’re able to understand where people are spending their time. We can get a whole bunch of hidden signals and data input from the the integration. So we can see things like how long someone was on calls for, their occupancy, alongside other data points to understand how they’re spending their day and what the challenges are. From a burnout perspective, when you’re dealing with increasingly complex call types or in the the high occupancy environments, The challenge tends to fall down to the front line. It’s your front line teams that are living the workforce challenges coming in every day to a, you know, environment that’s just call after call and you’re just kinda getting burnt out for throughout the day. The front lines are the ones who live that challenge, and then your management teams, your leaders tend to be the ones who pick up the pieces and try and solve the problem, from dancing between different dashboards to understand where the challenges are just to focusing where should I spend my time today. Should I focus on, say, Tristan versus maybe Josh, depending who’s having the the, bigger trouble. Outside of the actual hard dollar cost of recruitment, just the people cost, the people burnout cost is really, really, a big challenge that people are facing. And like I said, it kinds it it tends to land at the front line at the user level where where they’re cleaning up those problems. And and keep in mind, this isn’t just about reducing attrition. We’ll talk about what the product does in just a little bit and see kinda how we can help how we can help solve the issue. It’s also about making the people who are working at your companies ambassadors, feeling like they’re supported, that they’re really getting the attention that they deserve. So if they’re struggling, we need to do something about that and solve the problem. From a cost perspective, talking to a lot of different customers, it’s anywhere from twenty to thirty five thousand dollars just in recruitment costs. Sometimes that can include your ramp costs, including your your, increased handle time and time to rehire. So from a training perspective, how many months does it take for them to get up to full proficiency where they’re gonna, perform just like a regular contact center agent or someone who’s tenured who’s been there for a long time? On top of that, you have all of the people costs like I talked about where the environment’s kinda getting to you. So those challenges tend to add up, and and they they end up becoming extremely costly. Now I wanted to ask you guys if you can just drop the answer in the chat. How often are you talking about this problem really with your senior leadership teams, attrition or burnout, maybe the cost of recruitment, losing staff, things of that nature? Is it a monthly conversation? Is it a weekly conversation? Is it daily? I know that these kind of things tend to change depending on maybe budget seasons or however you’re balancing the the staffing. How often are you guys talking about it? Hey, Josh. There’s a lot of weeklies, a few monthlies, but the majority is weekly right now. Interesting. Okay. That’s awesome. It’s it’s it’s good to hear that we’re focusing on the problem. So from a a tooling perspective, the challenge tends to be figuring out what to do with the problem. Once you once you identify it, you can see signs of burnout and having conversations, but using data to understand it, that’s really where the big challenge is. That’s again where we come in. Sitting on this mountain of data, we have access to all of these different data points because of the ACD integration. So we can scoop up what’s going on in a contact center, and then we could start doing analysis to understand where are their patterns that we could identify that will maybe give us an indication, a prediction that there’s a problem coming, and then give us the tools to solve it, to do something about it. And that’s where the other half of of Intradaym’s offerings come in. So what we did is we we we’re speaking with our customers. We understood the whole problem of burnout, talked about the employee battery concept, and then Tristan and team, which we’ll talk about in just a second, some more details, basically stood up a machine learning team to get the data from those integrations, type it over to a place that we can look at it and start doing some analysis on it to understand where are these challenges and how can we operationalize this to solve the problem. Also, giving you the data to do something about it as well as the tools to take action and and solve it as it comes up. Tristan, I’m gonna hand over to you here. I’m gonna get you to walk us through the concept, please. Thank you, Josh. So as Josh just mentioned, we had a feeling that the gold mine for us was really the amount of data that we had from the ACD and WFM integrations. And so, the exercise of tackling the burnout problem within contact centers, started by discussing with SMEs, industry experts, and define hypothesis, which relates those data points to, burnout patterns that that are known in the industry and then validating those hypothesis, statistically, by running the data that we had, through some models and validating those hypothesis. Once they were validated, we were able to build a dataset, which was powerful enough and and, had a sufficient amount of data to train a machinery model. And so the machinery model that powers this agent burnout indicator solution is a classifier that classifies, daily each agent into one of four burnout risk categories. We have low risk, moderate risk, high risk, and critical. And so the assignments to categories are made are made based on the input features that are run through the model. This includes things like HT, different components of HT, so agent state information, occupancy. And really an important thing to note here is that, we could have built a model that has two categories, at risk or no risk. But the way in which we thought about it is having four categories allows to have a lot more continuous monitoring of the workforce and understand really where they lie within that burnout risk journey and be more proactive in, taking actions, implementing actions for agents by focusing on making sure that agents do not end up in the high and critical risk categories. So as soon as, we notice movement from the low to moderate and to high risk categories, action can be taken then. So very proactive, approach to to tackling burnout. Now the model was trained using those input features that that you see at the bottom, or some of them. We also have, combinations of those features. Those features across different time periods, are used by the model. And, we’ve demonstrated that our model achieves, up to eighty percent accuracy, on unseen data at identifying high attrition risk agents. If I could have the next slide, please, Josh. And so the this accuracy is we could achieve it in the way that we train the model, so using, again, this huge amount of data and also giving the model the ability to, to understand cyclical and seasonal patterns. We train the model on, close to one year of of data. And so, our model is able to understand weekly seasonality, weekly patterns, monthly, quarterly. All of those are are, known by the model, learned during training, and so predictions are made, in accordance with those. Now another important aspect of the way in which we build the model is, how it differentiates false positives and false negatives. In this exercise of identifying burnouts, of course, the the cost of making a false positive and the false negatives negative is is quite different. And so our model ensures that, it doesn’t miss any of the high or critical attrition risk agents and burnout risk agents, because the the cost of making a a false positive in this case is just providing additional support and training to, to an individual. So, it’s no real cost, actually. If I could have the next slide, please, Josh. Moving on to the dashboard, the actual solution itself. So, this is what a user would typically see when they open up the burnout risk dashboard. It’s organized in several sections. The main section, first one that opens up here, aims at really giving a direct quick insight into who are the most asterisk at risk agents and how is burnout risk distributed across the organization and over time. So that top table here, the agents that are at the highest risk would bubble up to the top of that table, and so a supervisor or manager would have direct view into who these individuals are. And then charts below this, this table and charts on the next slide also show how burnout is distributed amongst different business units, manager groups, supervisor groups, and over time. We also have a number of filters that you can see on the left of the screen, that allow a user to narrow down the search to only a couple of individuals, a set of individual, or just one particular agent of interest, and also time range filters, which allow to compare one time period with another, say, one week with the previous, one month with the previous month, or just go back in time and investigate a specific, time period. And the so that’s for the first section, really giving insight into where is burnout risk and how it’s distributed. And then the second section, of the dashboard, we call the supporting visualization section, and this shows this displays charts, that show performance metrics of agents over time and how those evolve over time. And the purpose really of this section is once a user, once a supervisor or manager has identified, burnout risks and wants to take action for those, this support supporting visualization section is what will, help for the decision making of the supervisor, and help the supervisor understand what action is best to take in in each individual situation, which might be different for, for two different agents. So really equipping supervisors with that extra set of eyes, right, with that lens to have them monitor the the the workforce, their teams, without having the ability most often to walk the floor, as could be done in the past, right, with with remote working now, is is really the whole purpose of that solution. Helping supervisors make more informed, better decisions as to what action should be taken, depending on depending on where an agent lies on their burnout risk journey. And so to support this further, we also have actually, an initial remedial action toolkit, that recommends for each, you know, each burnout risk category actions that could be taken, and I’ll hand it back over to to Josh to tell him more about those. Awesome. Thank you, Tristan. I learned something new every time you give me the the the presentation. That was awesome. From from an action toolkit perspective, this is really where our our initial offerings from Intradiem. If you’re familiar with what we do as a company, this is where a lot of these come in. Everything that you see within the orange lines, that’s something that you can automate. So these are actions that we can push into the Intradiem platform. We basically monitor the category or the risk category and becomes a flag that you can use to take action on. So as an example, if you’re working through the examples that Tristan talked about, those are more on the high end critical risk where someone needs to go and take action, really think about the problem and then decide maybe at a management level, are we gonna do some of those more serious actions? But, yeah, below that, once we get the logic in and understand where your categories are, we can start doing actions in the low and moderate risk category automatically. So from a training perspective, give them an extra training or a surprise break or just a a an extra break or recognition for their work anniversary. Small things like that that if you squeeze them into the queue while there’s already idle time, you’re harvesting idle time that’s sitting there. It’s productive hours that we’ve already paid for, essentially. There’s a lot of benefit on the engagement side of things, and it’s a way to take up some of that idle time and do something with it. So the this is really the idea is to use the dashboard to understand where the problems are and then eventually build the contact center ecosystem in a way where we’re taking action without needing a lot of manpower, a lot of work from your supervisor teams. And then if and when you do need to, we always bubble it up to the right person. So as an example, I wanted to ask before I go into explaining kind of what some of these solutions would look like. I have a couple of examples to show you folks. What actions are you taking today to manage burnout or to manage attrition? Do you do you guys have programs? We talked earlier about how often you’re talking about it. What kind of actions do you have? What kind of programs? Can you guys just kinda share some examples? Is it mostly focused around new hires and onboarding? Is it more for tenured folks who are kind of indicating those challenges? Is it a generalized project or program? Can you can you share, please? And and while we’re waiting for that, Larry, I don’t know if there’s any questions that I’ve bubbled up. We just got one question, Josh. How do you see this in action in a PBO situation where we, as the PBO, do not have the ability to manage to these items. I I’m assuming that means, like, a a business process outsorter or outsourcer or a vendor, like, a comp Yep. Yeah. Okay. Perfect. So so it’s essentially very similar to how we work when we have our our our initial offerings on a a BPO. It’s a conversation between the contract holder, whoever’s assigning it to the the vendor of the BPO, and with you you guys as the BPO doing the work. Where is the where does this value get saved? Essentially, there’s gonna be value on both sides. The person assigning the the contracts to your customer, they’re gonna want all these trainings and things to to be done because it saves money from a a delivery perspective. It makes you more efficient. But on the other side, from a BPO perspective, maybe you’re not paying for training hours. Maybe you do you you don’t pay for attrition because of the way that you structured your business. So it’s very subjective at the end of the day. We do have a couple of customers who are going down this road right now who are really, really heavy on the BPO engagement. And and our experiences, it it happens at the contract conversation at the the salary agreement that you have between the BPO and the customer. And it’s about making sure that if we’re getting value, where does it go to? Who who gets the money? Who gets the savings kind of thing? So if you want to, I’d love to have a further conversation on that to to see specifically what you’re thinking through, Dion. And, Josh, we got some answers coming into your question. No no formal program in WFM usually notified via ops, training possibilities, coaching, or inviting agents to be part in, for instance, a lean program. We we provide most agents forty five to sixty minutes of off queue time, four times a week, maintaining eighty five percent occupancy, balancing service level across several teams in time of need with cross training, focus on allowing off phone time in any potential pockets. Awesome. Awesome. Very, very, very helpful. And then just just as a as an ask, are any of you folks answering, do you have IntraDeum in place? Are you using automation to do any of that, or is that all kind of by hand? You’re you’re doing it with, with workforce folks kinda managing the buttons. Interesting. Yeah. We do have we do have a few using Intradiem that’s responded and then a few that are relying on WFM. Cool. Okay. I really appreciate the responses, everyone. It says awesome information. It’s really good. No. It’s it’s an option to use alongside c x one depending on what integrations you have, Heather. So if we wanted to dig further into that, we’d be happy to to double check. I wanna make sure versions and all that kind of fun stuff, but let let’s talk further. So so awesome. Very much appreciate the engagement, folks. So from an action toolkit perspective, like I mentioned, this is really the the playbook on how can we do something about it now that we’ve identified the problem. This is the remedy or the take action now portion of the conversation. Before we jump into that, Tristan, could you just high level kinda walk through what the, the the example that we went through when we launched this recently with the customer? And then I’ll talk through the actual solutions for the for the action toolkit. Sure. So we’ve we’ve launched this solution with, a major health care provider, in the US. And over, one year where we kind of piloted that, that solution, we’ve demonstrated very promising results for this solution and the actual, you know, value for the customer in terms of reducing attrition, reducing cost associated with attrition and agents burning out. So in the in the teams that, were using the, the solution compared to, the control group, We identified a seven percent reduction in attrition. That was seven percent, compared to the control group, five point five percent compared to, the historical baselines of of the team that was using the the solution. And, really, we we noticed great, value of, as I said, allowing managers and supervisors to take action early enough, and not only when they actually see all of the signs of burning out and attrition, But as soon as agents starting moving from low to moderate, taking action there, was key and really resulted in in, better retaining. So just a more tenured workforce, more engaged, better equipped, and and, just better customer service overall. Awesome. Thank you, Tristan. I love that example of being able to see, an agent or an employee going from one category to another. You can watch, see if the actions you’re taking are actually doing something about it. Great stuff. So as an example, like I mentioned before, this is the the lever to actually do something about it. So now that we’ve had the problem, I will just really high level talk about IntraDiem just as a platform. We don’t rip and replace anything contact center wise. We actually work alongside your WFM, your ACD systems. We can also have integrations to certain LMS systems and things like that, but we’re essentially complimentary to what the tech stack you already have in place. On the WFM side, we’re basically reading and writing schedules. And then on the ACD side, we’re getting data in real time, which enables us to identify all of these different data points that we’ve been talking about for the last twenty or so minutes. Once we have all that data, we can also see things like agent states or call states. So that lets us monitor for for problems that you would deal within a contact center in real time. For example, someone who’s stuck in ACW for a few minutes, traditionally, maybe you have a team of people watching a wallboard and they’ll reach out. We can automate that initial reach out. And what we notice is a lot of times, it’s just someone interrupted their day. Their dog walked in if their work work from home or their kid walked in, whatever the situation is, and a quick little reminder jumps them right back onto the phone. So you get a lot of self care out of things like that. And what that does is it frees up your your real time teams or your team managers or whoever’s doing that watching and support today’s world to be thinking strategy. Think about next week. Think about what your coaching plans are. Think about what collision is gonna happen next week because there’s a big call center event happening or there’s a big volume spike plan because the weather’s gonna be bad. Whatever it is that we should be focusing on instead of dealing with fires and running around to to do that firefighting, we can free you guys up so that you think of more about fire prevention. And it’s similar when you think about training delivery. Instead of kinda working through going back and forth from a training perspective and and prescheduling everything and then having to reschedule it when there’s a problem. We dynamically deliver training during idle time in the queue. So things like giving a recognition break or a coaching session or thirty minutes or fifteen minutes to go review training content, all of that can be automated using the platform. And like I said earlier, we could be used as kind of remedial action if you notice someone is moving towards a high burnout category or critical burnout category. So as an example, for the training side of things and we do have some video content that we can always share just kinda walking through what it looks like or what it feels like from an agent experience. But for for the logic, the longest call waiting and calls in queue, this comes out of the ACD. We’re measuring queue health here, Similar to what you would do if you were a team manager and you’re calling the real time depth to say, hey. Can I do some extra training or I wanna do a coaching session right now? We check the queue in real time every time we’re monitoring agent changes. So if someone goes from a call and they go into ACW and we wanna offer them a training session, we we can monitor the queue, check the agent’s schedule to see if they have a training assigned, see if they have a break coming up because we don’t wanna offer a break if there’s gonna be a training that collides with it. All of those really quick checks that you would do as an RTA team, we can do in real time because of the integrations that we have, and then we can deliver something like a training session. So here’s ten minutes to go and do this training content or a coaching session where the coach will go and sell and put themselves in available to coach, meaning that now I’m resource available for the queue. If the queue is green, the agent doesn’t have any priority training assigned, I can give them a coaching session. And, again, it’s a way to harvest that idle time and take it out of the queue without needing to do any action and especially without needing to call workforce or check wall boards or do any kind of calculations or tabulations. We automate all of that stuff. And everything that you see here, if there’s a workforce segment that gets written back, if we’re gonna update a coaching session or change a training session or or truncate things because they did the training faster than what we originally thought, All of those workforce updates are fully automated as well. And just to kinda close off on the concept here, from a training perspective or off phone engagement, it really is just idle time harvesting. All we’re doing is we’re taking we’re taking advantage of that time that’s sitting there. We like to think of it kinda like a time machine. Anything from e huddles where every week someone gets ten to fifteen minutes. I saw a kind of a development idea like that up in the chat earlier. Everyone gets x amount of minutes per week, and you make sure that that’s delivered while there’s quiet time in the queue. You can also do critical training where you can ratchet it up if it’s really important. If it’s a, let’s say, an iPhone launch or something that’s happening in a few days, we can really help to get these things out at scale, at speed without your workforce needing to have their hands on it because we automate kind of everything end to end on the planning, identifying, and then delivering on it. So a lot of words in a few minutes. Larry, did anything kinda pop up? I see a couple things here. We do not have any new, questions. Perfect. Alrighty. So now just to close out, we really appreciate everyone’s time. I really appreciate the engagement. That’s awesome information for us. We’re gonna go read through those and and and double click on them. If you’d like to, please contact us for a free burnout or attrition assessment. We’d love to sit down and kind of talk through the numbers. I talked at the beginning about kind of ramp costs and and how long it takes to get handle time cost back to what our regular production hours are. A lot of the savings we’re talking about on the initial slides where we talked about the cost of attrition, we’re really focused just on the cold recruitment costs. How much does it cost to post and to interview and to get them into their first seat? From there, there is additional cost when you have a new hire. You lose tribal knowledge. You lose some of that camaraderie and the engagement and and just, like I said, skill for how how quick they are. So we’d love to kind of talk further on that. There’s some contact details there. And if you’re gonna be in Nashville in a couple of months, weeks, two months, please come down and see us. We have a talk at eleven AM on how the right technology can enable a new approach to contact center leadership, and you can always find us on the floor and probably our bright green shirts, and we’d love to talk further. Alright, Josh. We had a question come in. Should the model also learn from more soft indicators, such as tone of voice, hesitation, and answering, maybe even verbal aggression, customer feedback, or first time right indicators? And if so, does it give significant results? Tristan, I’ll let you take that one because I know you guys did some deep digging on adding extra factors into the model. And Tristan, as as we talk about that, Nathan also asked about using, speech analytics into the model. Yes. Thank you. That’s that’s a great question. So our model currently is focused on ACD and WFM data. Does not include any, transcripts, call transcripts, or, any of those call level, data. Now we’ve considered, and we we will be considering, adding sentiment analysis from speech analytics indeed. So things like, not not the full call trend transcripts, but, data derived from those, is something we could in include in the model in the future. As of now, it it’s purely focusing on data from the ACD and WFM. So, performance metrics, indicators, and agent schedule information. Awesome. Thank you, Tristan. Alrighty. And just to close as a last question, I I’m not sure if we see it in the chat there, but if you guys could drop in what percentage attrition you guys are facing, it it ranges really depending on on how you measure it and what you’re looking at. We’re looking at what the total number is, anywhere from twenty to sixty to over a hundred percent in some groups. Just if you’d like to share, that’d be a great way to close things off. That’s interesting. That range that range is something we see a lot, Corey. But, Cindy, that’s interesting. The back office is not very high. I’d be curious to talk more about that if you’re gonna be in, in Nashville. I’d love to dig into it with you. Those are impressive numbers. I remember when I used to manage, it was somewhere nearing forty percent, and we were aiming to get it down to thirty five. And that was our big, hairy, audacious goal that we all aim for. So you guys are you guys are awesome. Doing very well. Yep. Thirteen percent looking there. That’s looking pretty good, Heather. And just for thirteen, twenty five percent. And just to clarify, are you guys all considering internal movement as well? Like, positive attrition is still the same problem. Right? Well, it’s good for the employee, and that’s what you want to to grow careers. If you lose a bum in the seat, you’re still losing a bum in the seat. Right? Right. Awesome. We used to call that good attrition, but it always it it really wasn’t always good for the workforce management team, but it but it was good for the company maybe not and for the agent, but not necessarily for the work force management team. That’s right. The the goodwill is worth the cost on that one. Right? Well, guys, thank you so much for this information. I know we’re all worried about burnout, all worried about attrition all the time, and it it makes such a huge impact on our, workforce management team every day. So we appreciate this information. Anything any last comments from anybody, on the panel or, from the Yeah. Becky, we got one last question. I think this will go to Tristan. What is the difference between, LLM and the hallucinations with, versus Intradiem solution? So difference really lies in the fact that our solution is, analytical AI, and, any large language model is is more generative AI. So there’s absolutely no risk of hallucination or any data leakage, data sharing, using our our solution. The the sole machine learning component, of our solution is that model that, I’ve told you about. So it’s it’s a closed task, really. It has a defined set of inputs and provides a deep defined set of outputs. So it’s it’s purely based on, yeah, mathematical rigor and and just, no risk of, it’s not generative AI, basically. So, yeah. Alright. Well, he does have the information up here on the screen about the SWPP annual conference, and Intradiem brings a huge team. So you will find them on the, sponsor showcase floor in their, hopefully bright green shirt, so you can’t miss them. And, they have a session at eleven AM on the first day on Tuesday right after the keynote. So hoping we’ll see a lot of you there. And, again, thanks to to Tristan and to Josh for the great information, and we look forward to seeing you in Nashville. Thanks so much, everyone. Have a great day.