Eliminating the Coordination Tax - Interview with Dan O’Connell of Front

0:00So welcome to the next edition of the Punk CX podcast. We have with us today Dan O 'Connell, who is the CEO of Front. Now, Dan and I have known each other for a while now. He was formerly at Dialpad and Dialpad, I think, were on the podcast. I want to say 10 years. Might not be that long, but it's coming up there. Well, I've got the gray hair to prove it. So, I mean, yeah. I have some coming in as well. Don't worry. Well, there you go. But, you know, we've known each other for a while and we catch up and talk about what's going on in the service and experience, sort of space and that's exactly what we're going to do today.

0:33And I've done a bit of work with kind of front or run some of their, their research that has come out that was recently, which we'll get into as well. So we'll kind of do a bit of a review of that, but Dan, hello, how are you doing? Welcome to the podcast. I'm doing great. We've got a dream force going on this week. So being based in San Francisco, it's a little hectic outside, but I can imagine I'm

0:59might know yourself or might not know yourself, or Front, can you give us a little bit of a background on yourself and also the work that you do at Front and what Front do, obviously? Yeah, perfect. So I'm Dan, CEO of Front, and then a background has been probably the past 12 years building software for customer-facing teams. I built a speech recognition company back in 2017 called Talk IQ. I was backed by Salesforce that had been acquired by a business called dial pad so i spent about seven years in the contact center space and in the past two and a half years have been fortunate to be the ceo of front and so we're a platform for b2b companies to manage their customers so we can be their agentic first touch point for customer support and then we can also be the handoff that platform for all of the handoffs and escalations and bring in their support and account management teams together awesome now like we sort of are i sort of alluded to at the beginning is that we've known each other for a while we catch up every and again have a bit of a rummage around it's it's always great to hear your perspective as you said you're in the epicenter of it right in san francisco dream force is going on and so what i wanted to do is actually just maybe explore a few themes i mean no real agenda other than just kind of get your take given that you're right in the heart of things in california where many of these new technological evolutions are taking kind of place and which is having profound effects on the whole customer experience and service kind of landscape.

2:27And I just wanted to start and get your perspective on, I mean, where do you think we're heading? Where's the kind of the customer experience landscape kind of heading? I mean, is it, are we all going to go head to the beach and have everybody, everything done for us or kind of what's, what's happening? Yeah, I don't think so. Okay. I think there's still very, very real, very real place for, for people around customer experience. And like, you try to have really profound statements. I think like it's very clear look at these agentic first experiences are going to do a lot of the remedial easy stuff.

2:58And I think they're quickly showing that they can increase in capabilities and do more complex tasks. But I think there's always going to be a place for human judgment, empathy. You know, there's going to be a time where you're so frustrated, you're going to want to talk to people. So I don't think people will get completely removed from the customer experience. And I think that ultimately comes down to a question for every company in terms of what are the experience that they want to provide. And so I think it depends on the type of business. Then I think like the more interesting stuff, I think that's like people all get that.

3:30And I think see that for the most part. And I think that there's two trends that I think about. And look, I'm always going to be a little bit biased because I think this informs like our direct product strategy. But I try to come on and actually be like, look, like here's some opinions and they're not just self-serving. Yeah. I do think, look, people are going to pick. And this is true, not just for CX, but I would say this is true for how people are picking software. I think people are going to pick fewer platforms. And so you can see that there's blending of teams or merging of teams. I think we see that trend in customer experience, customer support.

4:03You can see the blurring of lines between support and account management or customer success. You see the blurring of lines in EPD across engineering, product, and design. So I think one trend is people will pick fewer platforms to consolidate their teams to. And then two, I think the other big trend that I think we all see is it's very clear that I don't think there is one agent that does everything. So I don't believe in the super agent that's the only agent that you have show up. I also don't believe a world where there's suddenly tens of thousands of agents in every company that seems troubling, difficult to manage.

4:37But I do think that there's going to be more agents that show up. And so you're going to need to have a platform that can coordinate and provide oversight and government for those agents. And I think this is where you're going to pick fewer platforms. And then those platforms need to actually bring humans and people together. And those platforms all need to have oversight and governance for like the experiences. And so I think like those are the two big trends. And then the last one maybe I would add is I think it's clear that people want to bring their own agents to platforms, that not everyone is going to pick an agent that is from the platform provider and you see that we go home we play on uh anthrop you know with claude we play with open ai and you build things and you want to plug them into the platforms that you've selected and so i think that's an increasing trend as well okay so i said like i mean we've seen that happening increasingly over the last kind of like 18 months two years where people have gone uh we wanted to build a bit of a walled garden but we found out that we need to kick down the doors a little bit and because you just have to play with everybody else's in the playground And I think walled gardens work if you're the monopoly business.

5:45It's worked for Apple as a strategy. But I think there's questions whether that's a sustainable long-term strategy for many businesses. And I think what we have said, part of our intention, and we've got some interesting... You know, we did an announcement about this just a couple of weeks ago of this bring your own agent, this open platform, which is, look, we want people to use our native agents to do things. We recognize that people are going to use other vendors or agents from other vendors. They should plug them into front. And then if you're also building your own custom agent to do a specific workflow for all sorts of reasons why you would do that, then you should also bring that.

6:21And I think it's pretty limiting for platforms to say they won't allow that flexibility because I think there's like very real market demand and interest in having that flexibility. And so we want to lean into it and I'm sure other people will lean into it. And I think that's why like integrations and APIs are so fundamental to many businesses today. I think having multiple agents is like going to be fundamental. Yeah. I mean, I think the thing that's interesting about that, I agree with the point that there won't be a super agent. I mean, even three generative AI agents, the best bots were multi-agent bots, or they were architected in a way that there was multiple different functions that identified a tense and solved those that were crafted to solve particular problems, rather than just having this one super thing, which never really worked.

7:17I just think there's too much demand to try to do everything really well. Yeah. Right. And context becomes important. Like all the things we were kind of chatting about slightly, you know, just before helping on chatting is like all of that stuff. Like, I definitely think there's going to be this super agent that tries to do everything in its general purpose. But I think at the end of the day, it breaks down when you have specific tasks and specific things you need context and history. And so people are going to invest in these specific agents for this specific thing and then need to tie it into something bigger.

7:45And so, I mean, there's a link to all of that. I mean, there's two other kind of things that have emerged this year that impacts that sort of the broader SaaS enterprise tech space. You talked about this kind of people move to kind of platforms, you might see a consolidation. And we've seen some of that kind of happening already with people acquiring kind of things and building out capabilities and things. But there's two things I wanted to ask you about because these are the two things that have dominated that. One is on the vendor side and one is on the vendor and also the customer side, and that is this SaaSpocalypse thing.

8:20And then also the whole tokenomics issue about the cost of AI and whether it is really going to be cheaper to do it with technology than not. Is it going to be able to deliver on this promise? But let's handle the SaaSpocalypse thing now first, because I know that I wrote about it and you can help with some commentary on that. We talked about some of that before, but what's your opinion on the whole Sasspocalypse thing? I mean, are you going to be out of a job by the end, by Christmas? Yeah, hopefully not. And, you know, these are like the, they feel like the layup questions for me because obviously, again, little bias.

8:59But I think like what has played out is you suddenly realize, like, look, I think Sasspocalypse is overblown. I think there's very much real pressure and opportunity around, hey, do you price and package things differently in a future? where agents can do more work and we pay more attention to usage and then ultimately the outcomes, which we'll get to for tokenomics. But I think like what the past couple months, you can see this in earnings is distribution has suddenly become really important. Right. And I think you can see the large SaaS players, ServiceNow, Salesforce, for example, have had terrific earnings that those companies were written off for all intents and purposes.

9:40Right. Six months ago or a year ago. And I think it's become very clear that like the distribution is is a competitive advantage is important. And that those companies have demonstrated the ability to actually evolve and bring native AI experiences to their customers and that their customers are eager for that. And so I think it really demonstrates the strength of a lot of SaaS businesses. Now, look, I'm obviously biased in saying that. That doesn't mean that every business is going to survive. I think the right ones will figure this out and are demonstrating that. But I think like the initial reaction, a lot of the market and investors had towards traditional businesses of, hey, these companies won't be able to figure out AI and how to bring it to their customers in an effective way seems to not be playing out.

10:30Yeah. Or playing out differently. And so I think that for the right platforms, and I think we're a right platform, I think that's like that's a terrific opportunity. But it's not going to be given to us, right? It's not going to be given to any platform. And so that means that we have to be willing to disrupt ourselves. We've got to be willing to push on product velocity. We have to be really focused on delivering value to the customer and having them see it quickly because there's threats every day from a new agentic company that can show up. Yeah. I mean, I think the thing that when I wrote about it, and thanks again for the kind of for the help and commentary you gave me that, but what I felt like was happening was that it's the emergence of AI and particularly gems of AI and then all these genetic systems and the kind of increasing kind of like ever increasing kind of power of these kind of models and sophistication of them and the people were like oh it's going to like change everything and then it felt like well it was this massive hammer and everything looked like a nail yeah right and and you look back in history then there's been these things have happened before people like oh and then it then people adapt and you know evolve and it's actually not quite as simple as in straightforward and i think that yeah the thing that i kind of like came about is that there's i think there's there's platforms that are more or software solutions that are more general in nature and there's the ones that are more specialized in nature and i think that's what's it that's really interesting about kind of where how you guys are position because you're very much in the b2b sort of like space and you're dealing with that complexity that goes with b2b transactions and support and service and all that sort of like stuff and which is some super powered sort of ai model is not going to like solve that just just like that that requires kind of you know coordination and thought and kind of discipline and all these different sort of things and so i think that's what we're i think that's what we're going to we're going to see it's you know the technology is going to undoubtedly disrupt many players but and two things will happen one is that some people will be protected because of their their deep specialism in a particular domain kind of area and the second thing is that people will evolve yeah it's just like they'll just move yeah and i think to add on to that because i think you bring up like a really great point if your whole software is just doing one workflow and that workflow is relatively simple, then I think your business is probably at very real risk.

13:06And so that gets into both the complexity of workflows, the amount of workflows. Does that span multiple teams and systems that, again, we talk, we frame it as like this complexity. But I think that gives your platform stickiness. And again, then an opportunity to go and figure out what are these agentic experiences and where do they fit in. That, again, a simple point solution for one workflow that is highly automatable is going to have a much more existential threat towards them. But I think it's real. Like, I think the initial reaction was like very real of like every business is suddenly at threat and arguably existential threat.

13:43And so I just think the pace at which we expected that to show up and not believing that some of these large platforms can figure out how to deliver these great experiences is a little, as I said, like overblown. Yeah. And then circling back on the other kind of like looming cloud over this space as well. And it's been, I think, a real break on some people's thinking around what they do and how they do it is the whole tokenomics kind of issue. is people have gone this is brilliant we can we can do whatever we want and and then all of a sudden people go like oh no this was subsidized and the prices are going up and oh it's not as cheap as you thought it was going to be and and then all the kind of the layers start to get peeled off and it's still not completely clear kind of what it's all going to look like um yeah so how do you see that sort of playing out and and also particularly in addition rather is is that going to make people be a little bit more uh considerate in terms of where they apply frontier models and which version of them and do they bring their own and is it like an open source one or is it like an slm sort of thing in different parts of all of this and so are we actually getting to a point where we're going ah shock horror it's not as simple as we thought it was going to be we actually need to do our homework and actually kind of like configure a thing that actually makes sense for us how does that how do you think it's going to play out yeah i think like on the token token token token token token xp's i think look like we look at you know just being as a ceo like i pay attention to our own usage right across these platforms and can see every month goes up for sure and so there's the very real questions that that you have of like what is the output which is um to your point um fortunately like into shifting it to like us in the space we plan And I think when we talk and you spend a ton of time in, right, when we talk about customer service and customer support, customer experience, like different ways to phrase it.

15:41The nice part is like everything is measured and managed there. So people know their cost to serve. They know their time to resolution. They know their reflection rate. So out of all the places I would like to play and have can defend value, improve ROI, I think customer support is one of the few places where it is easier than others. Right. Okay. And I think we see that, right, when we get into code generation. Well, the vanity metric of, hey, how many lines of code did AI generate? Like, no one cares. Doesn't matter. I need to be able to measure product velocity. Are we increasing it? That's great.

16:19But at the end of the day, are we building things that people see a value in and provide a wonderful experience? That gets harder to measure. Sure. And so I think there's like a lot of spaces where the techonomics are like people lean into it and are like want to use these technologies because I think they're really enabling and great technologies. Truly, I think there's the very real pressure of like the ROI and, hey, does this really reduce my headcount and in what way? Right. Or drive some efficiency or productivity gain somewhere else. But bringing it back to the customer support, I think it's one of the areas where it's more easy, right, or easier for a business to go and say, hey, this is the positive impact that it is having on both the customer experience, but also perhaps the top and bottom line within the business.

17:03sure i mean do you think it also gets more complicated when people were initially kind of looking at it from a perspective of i.e maybe like it's a it's a kind of one use sort of thing whether it could be a chat bot that's a support bot or on the front end or it's like a coding application or it's it's a uh agent assist or whatever it might be and it's almost like a it's a single solution that can stand on its own and do us do do a particular job now when you get into more agentic type of approaches where you're trying to chain a lot of things together and possibly a lot of different agents together to to try and solve a particular kind of like problem or observing a process then the number the amount of sort of inference that might be you know needed to do to do that when you kind of you go you start with a query and then something goes on gone then it pushes on to here and then it goes back and checks it and then goes on for and then it you could get to a point where it's just gonna get larger because the kind of the level of inference that you have to control for across that piece of particular it's a multi multi multi-stage kind of process could be yeah you see you see me like smile and laugh for those for those that might be might be watching yeah i think it becomes like the importance of um you You know, it's companies like OpenRouter.

18:28So being able to route different, understand like different tasks have different needs. You might use those. You might not need the most powerful frontier model for a simple task. And so then you might use a mini model. And there's trade-offs to using mini models, right? Latency might be one of them, quality of output, another, sure you save some dollars in cost. So I think like understanding all of this, put pressure on the teams that are building of like the expertise. And then I also think like the thing we don't talk about a lot in customer support or I hear spoken about a lot is like you have a whole BPO business and category.

19:06And a lot of those are offshored and they're very you're you're talking dollars per hour. And that has empathy. And it's a human. So it's got empathy. All the things we talk about human empathy, human judgment. Right. You name it. And that's something that also sometimes is like missing, I think, from the whole picture, which is we focus so much on the model and we kind of ignore both the human experience. And we also forget about this whole existing category that is done pretty cost effectively and pretty good and exists today. And you've got to really drive the price down for that not to work.

19:43But I understand all of the benefit. We play, we are a customer service platform. So we believe in all of the benefits of AI. but I think this is like that natural tension that exists in the market. And I think sometimes it's just not even acknowledged or spoken about. Yeah. And I think that's absolutely right. And it also brings me, I mean, I think the interesting thing about BPO is that whether they play it right and many of them haven't and many of them have or are rather, is they understand that they can position themselves almost to be this center of excellence for using these systems and things.

20:18and therefore because they're learning across multiple kind of accounts and then that can again provide additional kind of value to kind of people that want to leverage them and so there is that kind of like there's not just a labor arbitrage opportunity there's a skills and experience arbitrage opportunity as well and so it's what i'm seeing the smartest kind of like people approach people do is just is actually trying to take a step back and really think about okay what is it we're actually trying to do? Because there is a danger, it seems to me, that this is it, right? It's like experience is supposed to be about differentiation.

20:57If the experience is the game and we're trying to differentiate on an experience, but everybody's trying to do the same thing. Yeah, you fast forward to the world where everybody's got the agents across every channel and that is the experience that you have. And suddenly, and if we're all on the same models, Doing the same thing, you know. Yeah. I don't think the world plays out that way. I think it's, you know, I hope, my hope is that not. And that's why I always go back and give the annoying like business school answer, which is like, it depends, right? I think it really forces companies to think about the customer experience they want to provide.

21:33And I think why I get really excited about our opportunity is, and we spoke a little bit about this too, is I think the handoffs become suddenly the customer experience. Meaning if you're stitching together multiple systems and the context breaks across whether it's agents, whether it's agents talking to agents and the context is lost or it's the agent escalating to the person and the context is lost. If that person has to start over and retell that story, that person is going to judge you, I think, more critically than they ever have in the past and get really frustrated. And I think the best businesses will be able to provide a seamless experience.

22:10They will come across as one company, one experience to that customer, even if it's across multiple systems and even if those systems span multiple people or multiple agents. Yeah. And I think that's like the challenge and opportunity. Sorry, yeah, challenge and opportunity. No, I think that's absolutely right. I mean, it's an old analogy that I use. I mean, say that this whole thing, great service, great experience is like a relay race. There's no point having the fastest runners if you can't get the baton right. yeah because that's the most that's the most dangerous the most critical kind of like in a bit your race is over if you drop the battle yeah and so i think that's the the thing you're you're you're pointing around the handover is like you'll need to that's where things break yep and i think that's where you where you see the stress and the pressure on like the systems right and the agentic experiences and i think you know we got i think everyone like i've it's funny i've taken screenshots over the past two and a half years, every day of kind of like every week of like different AI moments that have shown up.

23:11And I think we have been like so excited to like rush and apply this that sometimes we forget to take the step back and again, put the customer first. You can see it as like, we were so focused on like every billboard is around like deflection rates and around resolution rates. And nobody's talking about, hey, you can go deflect and resolve everything, but nobody's talking about the quality and nobody's talking about the customer experience, which is it doesn't matter if the customer is more pissed off to begin with. But every billboard a year ago did not have that. It was such a rush for, hey, we can automate away the people from the experience.

23:49And nobody was talking about the quality of that experience or really able to measure it from the customer point of view. And so I think that's all shifting. And I think it's been really interesting. We got new technologies and we run it. Things are all kind of slowing down first. I think that is fair. You're slightly older than I am, but we continue to make these same. Oh, crumb. We do this as humans every time. Our ability to not learn from history is quite profound. Is quite profound. But I think the thing is that also I know is that because I've done a bit of work with Zoom in the past. and you know yeah they are particularly interesting i think because they've gone for they've taken a more of a unified communications sort of like stack and i think as the the sophistication the maturity of these automated front ends kind of increase then it then the importance of great service and great experience is going to be connecting that sort of front end to the kind of the more complex stuff and having that unified site from maybe the contact center all the way through to the middle there, middle office and the back office.

24:59And I think that makes sense. And if all the data's all in the one place, it makes sense, easier to coordinate. But then you kind of like recently announced a new partnership with Zoom. And I was like, oh, that's interesting because I like the way that they're, what Zoom's kind of doing with integrating that stack and the automation on the kind of the front end. And then you're dealing with the complexity particularly in the B2B space. I'm a bit like, oh, Dan's coming on the podcast. Got to ask him about that. Tell me about that. I mean, that feels like a really, a really interesting partnership.

25:33How did that come about? Yeah, so we're super excited about that. And really it's about bringing Zoom contact center integration brings like inbound and outbound voice calls to friends, including SMS. And so we want to be the place that the customer facing teams are spending their time. We don't want people to go and tab switch again if we go back to how do you provide a seamless customer experience and so i think like that integration i'll say demonstrates like the importance of like technology partners but also people that are picking platforms to be their voice provider and saying how do we not if we're really happy with that how do we stay on that platform and plug it into the thing that is powering what we think is like the best customer experience yeah um and so that was really the intention around that partnership.

26:19Eric Yuan, the founder and CEO, is also separately an investor in front. So we've had a close relationship. But I think they're a terrific platform. We're super excited about that partnership. And again, it allows people to go and bring in a technology and a great platform they're already using and bring it to life within the platform that they're choosing to power their customer experience. And so we're super excited about that. and but and then the other thing i wanted to kind of like i mentioned it before is that and i just want to kind of like zoom in on the convey that's i have that problem the complexity thing that you're that you're tackling because yeah you did some research recently and there was a webinar you invited me onto the webinar to talk about it it's called the coordination kind of text and i think it's a really really interesting kind of thing because a lot of the topic gets dominated with technology right yeah and then there's stuff that happens outside of it and it's almost like real organizational life in many ways so give me a give me the headlines on the on the for the benefit of the audience give me the headlines on the on the research that you did and what what you found out and what it means yeah the coordination taxes it's all the you know and you said it it's all the opera it's like hidden operational costs that uh organizations pay to coordinate work.

27:36And so there's some really interesting finding. People can read the research at research.front.com, just a little plug on that. But what you find is like people are spending, or companies are spending three hours for every hour of work that actually gets done. And so you think of this - And when you say they're spending three hours, three hours doing what? Three hours trying to coordinate. So that might mean chatting with your friend across the way to get context to go back. It might be, I've got to go hop in a different system. So it's all of this hidden coordination work that nobody measures, right?

28:09It's difficult to measure. Teams don't think about it. The COO of the business doesn't think about it. And so like pretty profound stats that show up. And then you think about the world that we're heading, where there's going to be more agents. And so you think of, oh, this coordination tax is already a real burden and pressure on businesses. And then you think of a world where more agents show up and suddenly that the need to provide coordination across systems, across people, across agents is probably the most important, one of the most important things to get right in a scaling business. And so again, the coordination tax in this research like highlighted this across a number of different ways.

28:51And again, I think some pretty profound data that showed up from it. And I think what it also kind of points to as well is that there's some very real work that needs to get done in order to enable those agents that might be coming down the pipe to be able to do and to help out with that coordination. So it's a bit like the agents ain't just going to show up and magic's going to happen because there's a whole bunch of homework that needs to get done to be able to prepare the ground so that these agents can operate and then people can then respond more effectively and can answer problems more quickly and provide that better service and a better experience.

29:32particularly in that sort of B2B space where oftentimes you might be chasing somebody who's on site somewhere, kind of like, you know, who's got a piece of knowledge in their head or in their inbox or something. Yeah. And I think what's interesting, like, you know, some of the stats from it, like 42 % of companies don't pay attention to this. Like they don't measure it. And so then you think about, oh, you're picking a new platform, like slowing down, much like we said, like slowing down to speed up is understanding, do I have the right systems and processes in place? If I add more agents to this, is it going to increase the burden and tax that I have on coordinating things, or is it going to decrease it?

30:08And so again, I think the research is pretty profound on the amount of time people are spending behind the scenes coordinating work, as opposed to being able to just focus on customer problems. It was pretty surprising and shocking to see the number of people that don't pay attention to this today. So again, our goal across surveying 700 different customer service leaders was, I think, to bring that to life and show that this is a very real opportunity to help drive efficiency and a better experience within your company. Yeah, absolutely. And on research, I know that you're a bit of a fan of doing your own little timely polls, particularly on LinkedIn.

30:46And I saw one recently. It says, you'll be running a survey where early findings suggest that 75 % of CEOs and customer leaders expect support, customer success and account management to become more unified. But only 21 % say their technology supports cross-functional customer work extremely well today. Now, I saw that and I was a bit like, that's interesting because I thought my impression was that that's where more the technology is kind of going kind of faster than the leaders are. And I was a bit like, are you telling the truth? I think we're telling the truth. And these are like interesting things.

31:30Like, you know, what I like about my role is I had a different circle of people that I can access to. So if I ping somebody to take a survey and they're a founder, a CEO, or a chief customer officer, fortunate they'll like provide their insights. And so it was really interesting. And also to ping, people can find this on LinkedIn, pretty interesting companies that responded. So folks from like OpenAI and some really fast, innovative businesses. And I think like what to me is like what is coming clear is companies are going to need to stop organizing the customer work around internal team boundaries.