Start with Strategy, Iterate Like a Band - Interview with Craig Crisler of SupportNinja

0:00So welcome to the next edition of the Punk CX podcast. With me today I have Craig Crisler, CEO of Support Ninja and also a friend of the podcast. Craig, welcome, it says in air quotes, back to the podcast. How are you doing? Adrian, it's always good to see you again. Yes, long time listener, first time caller, not really, but second time, third time, we chat a lot. It's good to see you. Yeah, no, great to see you too. Ah, but do you remember when you were last on? oh shoot this was last year i guess no yeah yeah it wasn't it was it was october 2024 when the podcast went live wow yeah i know right it feels like it wasn't that long ago to be frank yeah everything's just going like kind of right now well now that i'm thinking about that conversation too how much has changed since then like literally like in the last two years it's massive amounts of change in our and in my space for sure and we've talked about how much has changed everybody's space to a certain degree no indeed indeed anyway um i mean you're obviously famous in your own lunchtime but can i ask you to uh introduce yourself and tell me a little bit about yourself for the people that don't know craig chrysler and don't know support ninja amazing so again my name is craig chrysler ceo co-founder of support ninja support ninja is a CX transformation firm, but we primarily work in the outsourcing space.

1:24We do a lot of work with small, medium-sized business and mid-market businesses, helping them on their CX journey and helping them navigate not only AI, but how and what ways you optimize your CX as you move forward. My background is I'm a startup guy, born and raised in Silicon Valley, startup guy since inception, 12 companies, four exits. I love working in tech. I thoroughly enjoy talking about technology and the role that technology plays in CX. And so always excited to kind of talk story with smart folks like you and about everything related to cx and on top of that avid music follower well this is the you know at the time of recording this is the coming out for the middle of may and we met for the first time face to face in seattle a couple of months ago which was great and you shouted me dinner which was even better so thank you very much i think we got the same because it was a bit like this big wrapped around sort of like fish kind of thing which was incredible it's like a red snapper or something yeah insult something which is ridiculous but brilliant yeah but i wanted to pick your brains a little bit today because you know you say that we kind of talk a lot and you've shared a couple things recently with me and i just wanted to that i thought were really interesting and i thought well it's let's kind of open the box a little bit on some of those things and the first thing that you said that that you talked about this industry or the industry that we're currently in is that an inflection point right now.

2:48And we look at things, we come at things from a slightly different perspective. We see these different things. And so I thought it would be really interested to get your perspective on that. And then, explain to me what you mean and what do you think is happening to force that inflection kind of point? Well, specifically because where we play is in the outsourcing space, or where I play is in the outsourcing space, specifically in the CX outsourcing space. So this kind of inflection of CX and outsourcing, which has kind of been, you know, tied forever, like they've always been tied together, kind of CX and outsourcing.

3:20And when you look at where outsourcing is and where CX is, it's been really this kind of, for lack of a better term, you think about headcount and tickets and cost reduction, like that's pretty much how people think about CX. And they've thought about it for years. It's a cost center yeah it also has a revenue center like this is the way it is and what's happened with ai which i jokingly say that outsourcing actually created traditional outsourcing created what ultimately will destroy it which is you know essentially that repetitive tasks being done over and over and over again by humans now can be done technically by ai like you know if it's well programmed and trained correctly can a lot of it can be done by ai and so now that model has basically been you know created and moved towards ai and so it's the model itself is outdated so people even the cx space they're starting to rethink how they think about volumes and tickets and stuff like that and they're moving away from this kind of like cost-centric you know orientation to being about like how do we think about cx more holistically because now we have tooling that can let us look at all of it retention customer experience in a life cycle now because you can analyze all that data more effectively you don't have to have a team of data engineers to do it you could do all this stuff differently now and so it's created this weird space where now it's like well what do we do now like because that model has changed fundamentally i mean i think it's it's interesting because i think there's i mean i've seen some interesting research i think it was Gartner that predicted that because the economics of, I mean, the potential to do what you're saying is absolutely there, but the economics haven't quite been worked out as yet.

5:03And Gartner kind of paradoxically is predicting that actually by 2030, it might be still cheaper to outsource some of your support function to nearshore or kind of far offshore than it will be to pirate by kind of ai so there's like still some things to be figured out it's not a one done sort of thing so it's like it's i think you're right as an inflection point but it's one of those inflection points where like it's not just we're going for one groove into another groove it's like we're we're changing but we don't quite know where we're gonna land right like they need to people are trying to navigate which way to go and like that's the inflection right they're like how do I do this differently now with AI in the picture?

5:48You know, and conceptually, it's interesting because actually I was just reading a report from Addressen where it said that CX hiring overseas has increased over the last two years. And so, you know, all of these like doom and gloom saying like, hey, we're going to lose all these headcount. Now it's like everyone's saying, well, technically, we're still hiring way more. And then actually on LinkedIn recently, there was this article that came out that said there's been more posts around customer support reps cx reps on linkedin than there's been in the last two years so okay it's interesting because like okay well this thing's happening that could impact what people are working on it but i think people are saying like what does that impact look like because that model itself is not quite making sense to them anymore because of what they can do yeah absolutely i mean and then within all of that because i think what's quite interesting is that it becomes like the AI, whether it's genetic AI or now kind of the genetic AI, it becomes always the next best thing that's going to solve everybody's problems and be the panacea and kind of make everything great again or make everything just great.

6:54But you also can, like I said, most CX problems aren't actually people problems. And I think I wanted to explore that because it's like shining a light on some of the things that people need to be thinking about and stepping back and getting away from the hype and all the maelstrom of news and everything and going, what is it that you actually want to do? And so I want to ask you to sort of, just to explain to me what you meant by most CX problems aren't actually people problems. Well, and to be specific, it's people problems in terms of the people who are doing the CX. It's not actually the customers.

7:31The customers are the people you're trying to sell for CX, right? I think that's what you're trying to figure out. Exactly. Customers are customers. they will be random and they will do whatever they want to do they'll be random and brilliant and you know chaotic and all and chaotic is it all probably in the same moment but yeah no so so the people's the more the internal people the people that are doing the the serving and the building and the delivering and things so right but tell me a bit more about where the actual problems actually sit well it's so even for us at support ninja like one of the things that we do a lot of is, so we have this transformation group and like their whole job is to look at the CX of our clients and try to understand what the problems are.

8:11And what we're finding in that group consistently is that a client will come to us and say like, oh, my average handle time or some arbitrary metric that they think is really, really important to their CX. They'll say, you know, it's not where I want it to be. I really want it to be at three minutes. It's at five. Like, you know, and they'll say, well, let's throw people at that, you know, and try to drive that AHT down. But then we go in and we look and you start to kind of, and this is where AI can do a lot of really interesting things. You start to look at the processes and the data and you start to unpack it.

8:41And then you realize, well, actually this is like an upstream problem, meaning that there's something in the production that's making this particular tool not work well. And now all of a sudden, it's not actually that you have an AHT problem. It's the fact that you have a product that's creating a CX problem, that if you fix that part, then all of a sudden that whole problem goes away. Or you're finding something in a return policy that's just a really bad return policy, but it's creating all these tickets and things that are coming to you as a CX professional. And you're saying, oh, we need to put more headcount to address with this bad return policy.

9:19Well, that's just poor CX. Go back to the return policy. Just look at the return policy and understand that that's actually driving problems. Let's tweak your return policy and make it better CX. And so it's almost like it's throwing bodies at a problem. You know what I mean? Instead of being like, okay, well, let's get down to the root cause. And when you start to look at the root cause stuff, what we're seeing consistently in the folks that we look at, very rarely is it that, oh, I need to put more people in it. It's more of like, let's automate this step or let's create more tooling around getting more information to the agent so they can have a more effective interaction with your customer or something like that.

9:57and not necessarily like, okay, let's add a thousand more headcount. Yeah, yeah. And I think it's, you know, what it's, I think it's, and this is probably relating, related to kind of what you kind of, the next thing I want to ask you about, which is probably a related point, because you said that AI is exposing and sometimes worsening CX kind of gaps. And I think that's, I'm guessing what you mean there is that actually you can't skip the fundamentals. Right? Right. And because it's almost just like a bigger, brighter light that sees into all the different sort of things. And if you speed something up, it's a bit like, you know, if you've got a leaky pipe and you put more water down the leaky pipe, guess what?

10:37The more water is going to come out of the leak. It's just physics. Exactly. It's wild to me because now what people are seeing, you know, in terms of now that you can look at the data in a different way and you can analyze it in a different way, more effectively and faster because now you have tooling that will allow you to do it you know you can turn on your clod and start pointing it to parts of your cx journey and then just ask it questions to say hey where is my problems and and it will find the leaky pipe like you know which you didn't know you had but yet you're still trying to push through that kind of old model of being like okay oh let's just take a volume and it's just that kind of monotonous monotonous work and i think it comes back to like the original like with every hyped technology there's the original hype you know which is and they put these like band-aid solutions on stuff so the super low quality chat bots that are out there you know there's answering the questions completely incorrectly and and creating more cx problems for you or you know the customers are frustrated around the fact that they really want to talk to a human but they're stuck in this ai voice thing and they're like how do i talk to a human can i just talk to a human please and and it's that band-aid is like not good like it's not a good band-aid but that's not really the problem the problem is actually in the internal organs like it's it's way it's way outside the cut it's like in the internal organs and you're actually having issues like within your systems and processes and data and and all that stuff that could really change your cx but they're just there's hot the hype is like let's get let's buy the band-aid and we'll buy the band-aid instead of being like oh well if I just dug into this and sort of dissect it a little bit oh I have this really bad kidney you know I have a kidney that's not good yeah I so I think the it's a an analogy as an analogy or a story whatever a metaphor it's a Aesop's fable is the race between the tortoise and the hare and it's something I wrote about a while ago and it felt to me that there was there was there was almost like everyone was getting carried away by the excitement and the hype trying to be always trying to be a hair but in actual fact they need to be tortoises because to create some of these changes it takes time and effort and you know deliberate sort of action and commitment and and all these different sort of things and and sometimes the um it's not sexy work sort of thing but like as a an old friend of the podcast from years and years ago this guy called Jerry McGovern once told me, he reckoned that 80 % of customer experience was cleaning and sweeping.

13:18I love that. Keep things tidy. Make sure they're all plugged in in the right place. Well, and it's interesting for us because in our work as a board ninja, we've had more instances recently when we're doing transformation with our clients where we're helping fix AI deployment that they've already done. And what that looks like is literally us going in and cleaning up their knowledge base and keeping it up to date and putting branding governance around it so the tool will work correctly and doing all of this not pretty work that's like very human, data-driven, you know, kind of quality type, you know, this is a bad answer or this is not true anymore and those kinds of things.

14:03And what's interesting is we can do it. It doesn't take forever, but we can do it. Now all of a sudden the tool works the way it's supposed to work. But they didn't want to do the grunt work, the cleaning and sweeping ahead of time to make it so that tool could actually do what it's supposed to do effectively. Because it can work effectively. You just have to do the front end work. Yeah. And it will work. But do you think it's fair that people have underestimated the amount of effort there is required when they adopt these new AI powered sort of platforms and tools in terms of the maintenance that is required?

14:38And that's one of the things that is catching them out. Because actually, they need to move to a new operating kind of model. Because that level of maintenance is shared between them, the vendor, but also it's now on them where it wasn't before. And that gap is something that's showing up and they're like, why are we seeing drift in the accuracy of these sort of things over a like time? time. And it's because all some of these things keep, you need constant development and monitoring and maintenance in order to, and refreshing of knowledge bases and all these other things in order to maintain the quality of the output.

15:10I mean, is that fair? Yes, it's 100 % fair. I mean, it's the, I mean, to use the jargon of the day, it's the human in the loop. You know, it's that, you know, if, you know, like any tool, and it's funny because IT professionals understand this, like server, there's like server maintenance people, like Like that their job was to look at the server, make sure that it's working correctly. They monitor it. You know, they do all this stuff, you know. And it's like any technology requires a human, you know, many times to be in there, checking quality, understanding, looking at error rates, understanding the knowledge base.

15:43And I mean, and it's interesting because if you think about an agent as a human, an agentic agent as a human or a chatbot as a human, like they're only working on the knowledge that you train it to do. And so if your product pivots, like you have a version 2.0, all that knowledge has to be updated. It doesn't magically know it, you know, even though they may, they don't read memos, internal memos, or it doesn't do any of that. So you need those humans to kind of be in there and looking at it, the data and saying like, hey, this knowledge is good. This knowledge is bad. Let's bet this out. Let's make sure this is getting trained.

16:16and what i've been particularly obsessed about lately is that you know any of these tools whether especially if external if it's external facing it has to have like like branding governance on it like you know and most people forget that and so if you have i always remember we have one client where we helped them and it was very much like this kind of surf sports gear type and their brand voice was very much like you know they were hey dude yeah exactly they wanted our agents to be like, hey, bro. Yeah, no, I feel you on that. That's an awful board. It's so bogus. Let's get you a new one. That was like their brand voice.

16:49That's what they wanted. And as we've been working with them, if you don't have that kind of governance framework to say, sound like a surfer, that kind of stuff, then the bot's going to come out and just say, oh, I'm sorry, your return didn't go well. It's not going to have the brand voice that you wanted. And from a CX perspective, that's not good. Then they'll know they're talking to a bot, which I guess you're kind of trying to avoid. Yeah. So the other thing you kind of said, because of all these kind of changes, that you also implied that there was a necessary kind of shift that needs to take kind of place.

17:26And I want to understand, is that like a mindset shift or is like an operational shift, a business model shift, commercial shift, or all of the above? I think specifically in the CX space, you know especially as it relates to outsourcing for us I think the the shift that needs to occur and it's interesting because I'm having this love-hate relationship right now there's this big kind of mantra within all businesses right now and I've heard it like so many times of that no more hiring more AI like no more hiring more AI and I don't love the term because it implies that it's that easy to do like you know that you're just not going to be able to hire more figure figure out if AI can do it.

18:06But it is that there's an underlying thing there of saying that like, hey, if you think about your work that you're doing and you approach repetitive tasks in a way that could potentially be reduced by AI, is there a world in the future that says, and this is where the necessary shift comes, is there a world where you say, oh, okay, if I look at the work that I'm doing, can I do more than being, you know, really higher value work, the stuff that really requires me to be involved? Can I do more of that high value work and then take the stuff that's repetitive off the table, you know, and then orient yourself towards that to be like, oh, there's stuff that I'm doing that I can basically take off the table.

18:48And so now I'm actually elevating my time to be doing really high impact work, high value work. I mean, tangibly for us, what that looks like is, so we have one client who we started to do analysis on their work. And we found out that our agents were spending 80 % of their time doing stuff that was just like repetitive, you know, password reset, shipping confirmation stuff, things like that, that really didn't require human thought. Like, you know, there's systems we can put in place, automation that we can put in place where, you know, you can get those answers faster. Yeah. But they're a very high touch product in terms of support because it's a little bit complex, it's technical in nature.

19:25And so our agents were getting bogged down in calls that just weren't important, not high value for them, you know? And what we found is that one, we didn't need to have more headcount. We can put some automation in the front to do that. And now all of a sudden their CSAT scores and their, you know, customer satisfaction scores were increasing because our agents then were spending more time on the phone doing really good work with the customer instead of doing that. Hey, did you get your box or not? You know, that stuff goes away. So that's the, I mean, I've theoretically, I've told people this, I'm probably one of the few guys in this outsourcing space that will say it like my, a lot of my job now when I talk to clients is not so much of being talking about headcounts, talking about CX and solutions and understanding that headcount is a by-product of that, not necessarily a necessity to perform that.

20:13And do you think that's, that's the, that's a gap that's emerging or missing kind of in the, in the market right now? And, and, and so what does the market do you think kind of need? I mean, what are you hearing from kind of clients? How does that affect, if you like, the outsourcing kind of operating model or offer, let's say, going forward? Well, it's interesting for us because I've been running up against more procurement managers that are stuck with like, oh, I just need to know the number of heads. Right. And we're saying, no, no, no, we're trying to solve a CX solution for you. Like, what is this problem that you're trying to solve?

20:49and let's dissect that before we even talk about headcount because we don't know exactly what we're getting into. And it's interesting because the market's still in that kind of traditional mindset of being like, oh, it's just the number of headcount. I need to do the basic math. I need to understand workforce management coverage and this is where I'm stuck. And where we're saying, and I think others, we're coming at it from a consultancy standpoint of being like, hey, there probably are issues that you have underneath the hood. can we ask you some more questions around it to see if there's a way we can get to a solution that makes more sense than just headcount.

21:25And that shift is rough in this space because CX has always kind of been seen as a cost center. And so procurement professionals and folks in the industry, they're just like, oh, of course it's FTE. That's what I need to understand. Why do you understand the number of bodies I'm going to throw at this problem? Instead of it being that there's a bigger solution around a combination of people, processes, systems, and data that allows you to kind of look at it holistically and say, oh, okay, well, maybe I don't need 50 headcount. I only need 20. And if we do this kind of work in the backend, maybe they'll be more optimized to improve my CX.

22:02Right. And does that mean that you end up, does your buyer then change? Do you have to kind of almost like involve keep your procurement involved but then procurement are procurement right and they're managing kind of contracts and costs and budgets and things um does that mean that you end up then having to go up or sideways or back or whatever to kind of try and make to do that extra effort to make that build that understanding to then come back to revisit sort of that and therefore But does it possibly extend sales cycles? It has for us. You know, it has definitely extended sales cycles for us because we come at it from a place of, you know, we believe, you know, CX is a holistic experience within the whole organization.

22:49Of course. Fundamentally where we sit like you believe too. And now because of the way you can look at everything and the way it plays together, it's much easier to have those conversations because we can say, hey, let us look, you know, and let us in and let us look a little bit. You know, we'll just, we'll work with you around it. And so those conversations do take longer because you're really trying to build a partnership and you're not just trying to build a body shop. Yeah. And it's confusing sometimes to buyers. There's also this like little hurdle we have to get over with our buyers because they're coming to us with a people problem at first, right?

23:22There was like, I just need people to answer these tickets. And we're saying, okay, we can do that. But just understand the background, we're going to be working on all this other stuff because we feel like you're not set up for success from what you're, otherwise you wouldn't have the problem necessarily. You know, and that's been, you know, it's definitely extending sales cycles, but it's also like almost like we're re-educating folks to a certain degree because I think they want it. Yeah. And the CX professionals want it. The directors of CX want this stuff. Like they see the future, they understand what could potentially be done and they want better CX.

23:55Like that's what they were doing, but they're being told by their higher ups of being like, you're a cost center. you just have to cut costs and meanwhile there's a way to do all of this and still cut costs like you know because it can be optimized yeah i think the thing that um strikes me i think i've advocated for this before is i mean it's based on that old kind of like idea it's like if you don't decide what you want you get where you get given type of idea right and i think the challenge with all this and it goes back to that sort of cleaning and sweeping sort of thing it's like and the kind of the tortoise and the hare sort of analogy is that almost like it feels like people that are service support leaders cx leaders whatever whatever their role they almost need to take a moment and go back to that what is it that we want to do and where and and and and and how does that apply to different customers across the different points in our sort of like our kind of journey.

24:58It's almost like get to that point where you almost have a service blueprint type of idea, where you go like, oh, here's a putative journey across different touch points. This is what we would need to deliver from a service perspective across that itself. Here's what the tech, human and tech balance would be, kind of hypothetically. And then to look at where we are now, how could we plot a path to get to those sort of things around, rather than just being this big bang transformation you go like i can rebuild towards that what would be the kind of the um the priority kind of points on the journey we need to achieve in order to get to get to that and that it feels like that often gets lost in it it's always that people can do it once and they feel like oh we know what we're doing you're like going yeah but what you realize you know what you're realizing not realizing is that everything's changing around you so fast that you almost have to revisit this kind of like on a frequent basis to make sure that you are you don't necessarily have to be bang on but you have to be directionally accurate to make sure that you're you know you're best best serving your your customers and your employees and and and the business with the choices that you're that you're making yeah it's it's that that education so like one of the things that we i mentioned that transformation group that we have that support ninja so one of the things we do is we put out this thing called a signal report so the signal report every client that comes into support ninja essentially what we do is we go into the back end of of our clients and we look at their cx journey and we understand where it is and we do a whole analysis compare them to their competitors to else work and this is all we just include it but we do all this work in the background and we've actually put them on a cx maturity matrix scale to say like hey you're here you know like you could be here but you're this like right around here right now based on what we've we can look at and what's interesting in that is most people they're like oh this is great because then they're like oh i've seen you look at this but we revisit it every quarter just to kind of see how they're performing against it just to understand what it is but it's us trying to train them to be like hey don't just think about this kind of triage work that you're doing which is essentially ticket management and how and what you're seeing every single day but realize that this is just a component of a larger cx thing that you're trying to create for your customers yeah and so if you look at it holistically take a step back and and then you look at it holistically because now you can more effectively and with more speed than you could previously now all of a sudden they're looking at it oh okay well that is interesting if i did this this could be potentially be the outcome it's like yeah you know this could be where you could optimize towards yeah but you're right i think you know it's like it's that whole what's that old saying you know what gets you here won't get you there Like, you know, it can get you to a point and you can stay there if you want.

27:46But if you're really thinking about how much is changing and how much the world is kind of, you know, technology is evolving, customers are evolving, you know, new generations of buyers are coming and they operate completely differently than the last generation. You know, you have to kind of revisit it over and over again to understand what it is. Like I'm fully like really trying to figure out right now. I have one of our COO, Jake, he has a couple of new kids and they're AI natives, right? Okay. They grew up in a world, in a native AI world, meaning like they're going to grow up with AI. And I'm really fascinated to see what they'll be like as buyers.

28:22Right. Okay. You know, because like, how will they act? Like, will they be driven by what AI is telling them or will it not be? like what you know like i don't know any of that it's so that part's intriguing to me but those people will buy completely different than my 21 year old yeah yeah you know because of this is just a different world yeah and so um so thank you for that so there was one thing i also i wanted to ask and this wasn't a thing that that that that you raised but it's something that had been occurring to occurring to me and i wanted to sort of pick your brains on this because as you said there's a if ai is going to get implemented let's say it gets implemented well and it will continue to progress and mature it will reduce and automate a lot of the repetitive and high volume going to work and then agents will end up dealing with more complex more urgent more complicated sort of like stuff so here's my question and so that and there's because it's more about those agents and particularly new agents and then sometimes more experienced agents when you can do automate some of that simple stuff because it makes me think about we know that contact center work or service and support work can be challenging work and it can suffer from a degree of churn you know and you get turnover people and there's only there's always people that are coming into the kind of the industries but if you're if we're automating a lot of that simple stuff then here's my question how do people learn and then the second part of the question is like also traditionally or historically many more experienced agents yes they enjoy some of these kind of like the more complicated queries that they can get their teeth stuck into but they've also i understand that they also kind of go and choose when they've done a series on them they might choose a simple one just to take their foot off the gas for a little for a little while just to take a bit of a breather because you can't sprint a marathon right because that's just kind of kind of fuel kind of burnout and things so there's two sort of two things it's like how do we compensate or how do we how do well how do new agents learn and then two how do we look after the well-being of experienced agents when there's no place to for them to take their pedal off the metal as it were right right because if they're only handling the most intense calls then And how do they navigate that?

30:45Well, I think like on the learning thing, I mean, I can tell you how we approach it at Support Ninja. Because I think one thing that's really big for us is that, so kind of what we were talking about before. So there's this whole human in the loop aspect of this work now. And essentially how that looks in the, I think how this looks in the future and where we've seen it work with a couple of our clients where like we have one client where they're basically like a pass company. Like they basically give out passes or a corporate past company and so you can go get yoga and all this stuff and so what's interesting about how those agents are working now because there has been a deploy you know like this kind of you know wide scale chatbot solution for the end user customer and so it's really reduced the number of tickets that you know technically these incoming new agents are doing and what we end up doing is more of those agents now they're incoming and doing they're actually the ones that are back in of that model, you know, that are training that model, watching it behave and, and, and making sure that it's of the quality that needs to be, because you still need those agents doing them.

31:47They're not disappearing completely. It's not, you don't need as many, but like, you know, and they're looking at, and that's actually how they're learning. Right. Okay. You're actually QAing the bot that's using the knowledge base and all the other things and the inputs that we're giving to it. And so now they're learning like that kind of foundational knowledge to understand how this company works. And then what's ending up happening is, and they're getting moved up into like the higher level as they get more and more skilled. And so it's almost like the entry level becomes an SME, like, you know, it becomes like a knowledge management person.

32:20Okay. Which I think is a very intriguing model to me because now all of a sudden you're, you know, technically training them in the tech with that ultimately that they're going to end up being utilizing in the future. the other aspect of it is there's a support part of it where for us like every agent in our company learns ai like gets prompt like they call it prompt like a ninja training like they everyone goes through this so they understand the kind of conceptual aspects of what's there because that's the future of work that's how we'll be interacting with a partner here at chatbot like to talk about think about and interact with things so i think the model of learning has changed a bit because now we're talking about discernment and less about knowledge retention.

33:05Yeah. Which is definitely different than the old model, which is seriously like how quickly can you memorize this script? Yeah, well, indeed, indeed. And that was just something that made me kind of think about it. But what about the experience agents and things and being able to like take a step off and stuff? And do you think that's going to be solved just by greater empowerment or people just to go, oh, I'm just, you know, when I need to take a step back, I'll just take a step back. Theoretically, you know, I'll be honest and say, I haven't seen this play out yet, 100 % the way I hope it will.

33:41But theoretically, the way I see it playing out is that, and we've seen it with a kind of a little bit of a couple accounts where technically, if you think about it, ticket volume is just going to naturally decrease if you have kind of an optimized solution. So, you know, you're just not going to have as many tickets bubbling up because also too feedback's going to engineering taking care of tickets the way it's supposed to and so what i've seen happen is now like they're getting tickets but it's just not at a pace that you've had before and so they're they're not like back to back against tickets and we're not trying to staff to a workforce management model is doing we're trying to optimize towards that interaction so if it's a 20 minute call or a five minute call doesn't really matter because the volume we're not thinking about in terms of volume metrics right okay And so I do think what will happen is, is it will get going back to that concept of these being knowledge workers, like it'll be more of a knowledge worker model where there'll be like, well, what it's a reasonable amount of volume that we think we can have.

34:39You know, what's a reasonable amount of volume that we can ultimately get to that will optimize the CX and that particular level of interaction. And I'll give you an example of one way it's kind of playing out. So we had a customer who came to us and it's a GLP one company, you know, super hot, you know, very big market. It's got growing like, you know, crazy. And they really leaned into the fact that they really wanted the interaction with a human to be highly impactful because this is like a personal health decision. Right. So it requires a lot of feedback and there's a lot of moving parts, doctors, pharmacies, the whole nine yards.