This week, hosts Michelle Schroeder (SVP, Marketing) and Nikola Mrkšić (CEO, Co-Founder) of PolyAI talk about the crux of boosting both operational efficiency and customer experience. The hosts examine how the pitfalls of traditional interactive voice response (IVR) systems have shaped customer perceptions, and how a blend of AI and human interaction can improve the overall contact center experience. Also, hear some amazing use case stories shared at the first annual PolyAI VOX event: https://vox.poly.ai. 00:17 Recap of PolyVOX Event 00:55 Discussion on Efficiency vs Strategy in Customer Experience 01:17 The Challenges of Contact Center Leaders 02:01 The Impact of Operational Efficiency on Customer Experience 04:15 The Role of Innovation in Customer Experience 05:49 The Evolution of Key Performance Indicators (KPIs) 08:07 The Intersection of Operational Efficiency and Customer Experience 09:25 The Role of Different Stakeholders in the Buying Decision 14:23 The Future of Customer Experience with AI 16:56 The Role of AI in Bridging Operational Efficiency and Customer Experience 20:19 The Impact of Previous Technologies on Current AI Adoption 21:41 The Role of AI in Contact Centers Follow PolyAI on LinkedIn Watch this and other episodes of the Deep Learning pod on YouTube
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
MichelleMaybe your CX leader, the person who owns, the lifetime value owns revenue numbers associated with a customer would be the person to say. I want poly AI, for example and I want the data that comes with it. That's not to say an operations person wouldn't but the goal of an operational efficiency model is probably not to close that loop in acquiring customers and then getting all the way to evangelizing them.
MichelleHello everyone and welcome to Deep Learning with Poly AI. I am one of your hosts, Michelle Schroeder. I'm here with Nikola Mircic and we just got off of our very first annual PolyVOX our first annual event for 2023. It was so incredible. Nikola, how are you feeling about it?
NikolaOh, amazing. It was really good. I think we've had a thousand people, a lot of our existing prospective customers, we talked about voice, where it's going. Show people a lot of the generative stuff that our product team has been working on really hard. So I'm really pumped. I'm excited.
MichelleYeah, I am too. I'm pumped, but also the adrenaline is starting to go down. We've all been up since
Nikolathe crash is coming,
Michellethe crash is definitely coming, but not before we have a discussion about efficiency versus strategy or really, operations like operational efficiency and customer experience these two competing interests. And for your opening session at Vox, you talked a lot about this, you opened up with Really a lot of the frustrations that sort of drove you to start poly. I the sort of bad calls, the bad experiences with technology that are really an extension of the tough spot that contact center leaders have been in. Probably since inception, since the beginning of contact centers, which is just, how do you meet more of these KPIs?
MichelleHow do you address more customer concerns? How do you just do more and more until you've squeezed out the sponge and left, nothing but a hollow, lightweight kind of customer experience behind. Yeah, yeah. Talk to me about that, a little bit, like you, you talked about the customer sort of being stuck in the middle of this experience between these two interests.
NikolafoR sure. Look, I think that when it comes to customer service, it's a cyclical business and this depends on the vertical and the shape of the economy, but whenever companies are in a tough spot and they just don't have an appropriately staffed and trained contact center, they resort to the dark arts which is well, the best, or maybe the most expedient fastest to apply form of automation available at that point in time. So I once upon a time spoke with a very large US carrier and they told me very proudly in a team full of IVR engineers or double my age that They're really good at this.
NikolaThey don't need help. Their containment rates are 92 percent and we do a lot of work I think we have a state of the art conversational system for getting both containment and a good experience in 92% Unless you're doing something really simple, like answering questions, maybe, or, like front desk for hotels is one of our applications where we get that kind of performance. But if we're talking about complex billing issues, if you're walking customers off a cliff, trying to get them to not churn, I'm like 92%. These guys are either running a secret Manhattan project, a conversational AI, or, and then I dialed into their system.
NikolaIt's an aviary tree that is. a labyrinth of, epic proportions. You just can't get out. So yeah, you are very effectively contained as is any shred of dignity or a good experience that you might've hoped to have when calling that phone number. And lo and behold, trust pilot score 1. 0 out of five, not even 1. 1. So I think, this company didn't do this because They hate their customers. They just had to at some point, they couldn't pay for it. The market was what it is and that, contact center manager at one point said, do it, whatever, you have a cramp in your leg let's amputate the whole thing and then you won't have a cramp in your leg, right?
NikolaSo it's definitely not something that long term generates value for the company, but these are the compromises people made and then we got stuck there. That's it. That's the efficiency side. Yeah. It can be quite efficient.
MichelleYeah. No, it's interesting. I think a lot of times, like when investors are looking at, stocks or like public companies and they're thinking like, what's the profile of a company that I believe is going to be healthy longterm, right there. It's not really like they're looking at things the way that we're, we are. Right now, we're talking about sort of operations versus, customer experience, strategy, but they are looking at something which is around like profitable growth versus kind of innovation. So like expenses and being able to keep that under control, operational efficiency versus their ability to innovate and keep up.
MichelleAnd what I thought was so interesting is you talk about the IVR. Engineers that were twice your age and saying no, we're good. We don't need anything new. Like we've got our KPIs, the KPIs that matter to us check. But I think it's partially because at that time. And I don't know exactly when this was, but I'm assuming it was pre COVID. Customer experience really was not considered an innovation. It wasn't, that's just not really how people thought of it because it wasn't even necessarily like a technical challenge where you need to have all of your stuff in place, like the digital transformations, you needed to understand how to meet your customer and customer data and react to that and actually make a better customer experience based off of what you know.
MichelleAnd because it wasn't. Always considered an innovation or an innovative thing to be a customer experience, first company. I wonder if the KPIs weren't wrong, containment, which sounds like jail the way that you described it. Do you agree with that? Have the KPIs changed since the beginning of you starting this company versus now?
NikolaAbsolutely. I think that the one thing that has been happening, I think probably for a decade now is that companies are getting better at knowing what one point of NPS is worth for top line and for their lifetime value of their customers. And these are difficult problems to work on and to attribute, right? Because. You could say anything. And most parts of the company are incentivized to exaggerate their contribution to this. And it's just a, it's not an exact science. But, I was in a meeting with a large financial institution a few months ago and they had a very precise number measured in millions around like what one point of mps meant.
NikolaAnd I was like, wow, cool. This is it like they know, because it's always hard to write a good business case for that improvement in customer experience, because the shorter waiting times mean that, it really depends on how your competition is doing. The reason I said initially that it really depends on the vertical and, geography is if the whole industry has converged. Converged on long menu DTMF where you just can't get out. If all the telcos are behaving that way that's just okay then, right? You're not really losing customers because you have it.
NikolaNow, if one of them upsets the status quo, that's where things get really interesting because then they have a problem and they need to react to at least have parity with those other companies. And when we look at our customer base a lot of our buyers are people who never wanted to deploy that kind of experience. And, we pride ourselves on the fact that, a company like say Caesars works with us when they've never worked with previous generations of, this technology, because they are just, they were unwilling to take that hit to customer experience.
NikolaAnd I think, some metrics aside, I think it was clear to all that you're not making your experience better when you put these things in place. Efficiency versus quality of experience, they were making a pretty conscious trade out there. Now the real challenge for our generation of companies is. To provide both, and then it's just a trade off, right? These can probably be at forcing extremely strong containment. But the good news is they're way better at balancing the 2 things. Yeah, so that you can get something that picks up all the time and just gives enough time back to the contact center.
NikolaSo that what's left of it can do what it does. Like that human part, things that are not maybe transactional things that are edge cases, extremely high value customers that you really still want to keep, assigned to your best agents while, someone's updating an address and our voice assistant works perfectly. Frankly, it probably works faster than a human because. It actually has a higher a lower error rate at transcribing numbers and zip codes and things like that than a human.
MichelleYeah. I have a controversial question that you just sparked, which is just the people who own Certain types of KPIs. So it's awesome that large financial institution, they had a sense of what a point in NPS is actually worth, but generally I don't know how many people within an organization are drawing a direct line between the customer experience in the contact center. And the revenue, average revenue, like per customer. And I wonder if not just, we're talking at a very high level about sort of operational efficiency versus experience, but those are often different people within an organization too.
MichelleSo you might have like somebody who's in charge of operational efficiency or an operational leader versus a customer experience leader, somebody whose sole job it is to make that CSAT score go up and make it mean something in terms of revenue. And I don't know if they're at odds with each other. I don't know if they're both making buying decisions, but I would assume that the operational person bought the IVR, back in the day, like back in, in this version of enterprises making that decision to like lower the cost, but make sure those calls are getting answered.
MichelleThey would make that type of decision. Whereas I'm assuming like an experienced leader would want to make a different kind of choice, maybe a more expensive choice. And I have examples from my past life and other companies, but just curious about that. Do you see those 2, not just competing interests, but competing individuals within an organization? And do you think that's a sign of the maturity of that organization if they are still sitting far apart?
NikolaI think it just happens organically, right? And we've had cases where we work with companies where customer experience is just trying to make sure that this thing that someone else bought doesn't impact customer experience negatively, right? And they're not really marrying the kind of like the objectives of what we're doing with the overall business objectives, but they're looking just at like, how do I make sure that this has the least? That it impacts customer experience and, or just changes the experience in the smallest way possible.
NikolaAnd that becomes problematic, right? Cause then design choices start getting controversial instead of having a voice assistant that really leads the way that a human would, that tries to create the most expedient, fast, good experience the way that the best of the human agents would They're like offering you to speak to a human if you ever want their disclaimers, they're apologizing. And guess what? If you're a voice assistant, and the first thing you do is, Hey, here's an answer to your question. Do you want to speak to a human? People are just going to say yes. So your containment slow.
NikolaAnd frankly, you've just added a completely unnecessary barrier to speaking to humans, which is what you're going to get a 95 percent of the cases anyway. So in that case, you paid poly for that technology. You've put an annoying barrier in between the contact center and the consumer. And it's just, it's not really going to work that well. Conversely, if you. Really then like optimize for a thing that would engage in a way that a human would then you create a bit of friction, right?
NikolaA bit of positive friction where you try to get the caller to engage. So a good example would be what do you do when someone stays silent, right? Staying silent is one of the most popular ways of getting out of a conversation with a voice assistant, right? Hey, didn't really get what you're saying. You're waiting. You're like, Hey, sorry. Could you say that again or whatever? Maybe you're hearing just like loud breathing or something like that. At that point, what we recommend to most of our customers is hang up. A, it might just be a connectivity issue.
NikolaB, let them say something, right? Cause at the very least, the best thing that a voice assistant can do is, they don't want to speak to it, right? Who can I route you to, right? Like they do need to get there. So people are a bit more used to that first generation of IVR, which is used for routing more so than for like full on task completion. They say something and you go Oh, is it about that? You lead proactively with the most comical type. They're like, yeah, then you go a bit further. Maybe if you haven't automated the whole thing, at least what you've done for that specific consumer is you've brought them a bit closer to a future in which interacting with a voice assistant.
NikolaIs as natural as dealing with a smartphone app or a good website. And that's what we have to work on. And yeah, is it a sign of maturity of those organizations? I don't know i'd blame us the vendors and our maturity and what we've achieved where you know They just haven't seen it happen in practice enough if someone's like coughing up, a six figure commit to implement this kind of technology They probably thought it through and have some metrics. So then over time, what we tend to see is a convergence between, COO and the CIO, and just, customer experience, contact center around metrics that are the best representation of what those companies are focused on.
NikolaAnd that depends on the company, right? Some never want to miss a call. Others are really looking to. protect themselves from having to, pay a BPO for an extra 500 people over Christmas if there's a retail company and have big beaks and throats. So yeah, I think it's it's a bit more nuanced than that, but these guys know what they're doing. I think it's the faith and the technology and the vendor's ability to deliver it. That's been like, that's why they're stuck. And, Those IVR conversations where they think that nothing better can be done.
MichelleThat makes a lot of sense. I wondered too that the vendors maybe have blurred those lines a little bit or, but I wonder if a few years from now when we answer and we will answer in this podcast the question can you actually do both? Can you actually be both? As a brand or even as a voice assistant, can you support both operational efficiencies and customer experience? Like really? But maybe in a world where the answer to that question is just an unequivocal yes. And you're like, yeah, absolutely. You're supporting both. Maybe the buyer profile does change.
MichelleMaybe your CX leader, the person who owns, the lifetime value owns revenue numbers associated with a customer would be the person to say. I want poly AI, for example and I want the data that comes with it. And I'm saying this because I was speaking with someone, a prospect at the time, a client now who is basically saying that she wishes that the insights that came out of the call center fed their creative agency, because they're getting so much information from what clients actually want, what they care about that it makes more sense to use that data and use those experiences to attract.
MichelleOther clients that are like them and it's just dawned on me, this is exactly what an experienced leader, how they would be thinking about closing the loop. That's not to say an operations person wouldn't but it's not the goal of an operational efficiency model is probably not to close that loop in acquiring customers and then getting all the way to evangelizing them. So it's just an interesting thing to think about. Anyway, let's get back to the actual question. The podcast is, do you think we can do both?
NikolatHat's like what we're trying to do, right? I think what you said is really interesting because the personas, they vary across different industries, but often the contact center is like has natural allies, right? So if they're taking, a lot of orders, if they're taking appointments and stuff, they're probably pretty close to the commercial part of the business, right? And that may not be the same thing as an operational. Side of the business, right? Often we have, and some of our best buyers are people who are in marketing organizations because they have the data, they love getting the data back and they have the data around like what this means for the top line.
NikolaSo it may not be like that NPS point vis a vis churn and, long term lifetime value. It may just be like very flat mapping from Hey, if these guys do better on that SLA. We get more bookings, right? So we often have like marketing there. And I think it really speaks to who is the buyer for AI in general, not even for just our product. And, it's an executive who knows how to translate its value into well, those use cases into value for their own company.
NikolaAnd I think that kind of talent merges organically across. Companies, it could be in, CX, but it's often in marketing. It's often an ops CIOs, especially more commercially oriented. Ones are often in a really good place to do this and to bring the business and it together around this. That also depends on the delivery model, but yeah, I think we can have it all. I think, once upon a time people were told go online to book, and that was just like online to book your airline ticket.
NikolaAnd, I remember, the Belgrade of my youth and you had like Yugoslav aero transport boots everywhere. And they're very nice and shiny and you would go and you'd buy a ticket. And I don't think there's a single one left at this point. And it's hard to imagine, if a website for an airline is a bit clunky, hard to use, you're probably booking the other flight. If you have a choice. So I think there, you go through like a drop and the good news is the drop would voice assistance. That's already happened, but all those screaming calls, people screaming for a representative, hanging up, staying silent to, to navigate that, like we've hit rock bottom, right?
NikolaAnd that's not to criticize the first generation of the vendors, but it's just. How technology goes, right? And I think what's always been fascinating for us as an Anglo American company that has, deployments in more than 20 countries. We probably have the first voice assistant ever in Poland and Sweden and Croatia. People engage. Much more freely. And they just skipped a generation of these things. They have better conversations, right? It's the Americans that we have to walk off a ledge and get them to engage through trickery through pleading, begging and we're changing their mind as well, but, they've been burned before and I think for a lot of these other places, which, they've never had a voice activated automation.
NikolaLike they're just sailing into a world where this stuff does do both. And I think that's exciting.
MichelleYeah, I agree. I agree. And not to sound too much like a Pollyanna, but I think that IVRs and like all of those bad experiences, they really paved the way for tools like ours in a lot of ways, when. IVRs came out and the contact center leader who is constantly being asked the question, how can we make this cheaper? How can we resolve more of these calls? They got the IVR sort of technology. The, it became a very obvious necessary disruption. It was, it became a trade off that everyone was very willing to take.
MichelleAnd I think, yes, it caused a lot of pain. It caused a lot of customers to be really unhappy with the phone channel. But the fact that we were already starting from rock bottom meant that the voice assistant, the conversational AI, industry had some room to experiment. They had some room to try things not to plug Vox or our annual event too much, but what I found just stunning as as a person who was like observing a lot of what was happening was all of our clients were Not just early adopters. They really blended these two ideas together.
MichelleThey really they saw the vision years before. I think the world is seeing it now. Which is just yeah, maybe we can have both. This is how I'm planning on managing expectations, and the hiccups along the way. And then, I think if you heard just the overwhelming positivity from the nine clients we had on stage today it just shows you, like how far we've come from that.
NikolaNo, for sure. I felt really warm and fuzzy listening to a lot of it. I didn't even know some of the best data points we heard, right? There was someone who just said I would have had to hire 200 extra people. Oh, yeah. It's impossible in this climate... when we look at this technology and does it offer the right trade offs people forget that for large contact centers, a lot of it historically has been, someone managing their own internal agents. And then a number of different BPOs where they're dealing with probably like partners that work with for many years, like large BPOs, like teleperformance, web help, Capita.
NikolaAnd they're used to like saying, Hey, data stayed there. Their agent churn is not that bad, so they'll get better in six months. These guys here, they're cheaper, but they churn a lot of people. They're, not very good at supporting me when the SLAs get tougher. So they're used to like managing those trade offs already. And you could look at an AI solution as, flexible, moldable, supply of agents of a certain quality. And the fun part is, they're different in some ways, right?
NikolaSome things are really good in other ways. They. Are not very good, and they're improving. But overall it's an optimization problem that especially at the larger side of enterprise, these guys know how to do it. They know how to do it well. I think that they've been trying to get that best mix, they're handed a budget and they're like I can't do that. I can't hire people here in San Francisco. I'm going to go higher in Manila, where I'll have more agents so I can get that, costing down. Cause at the day, they're P and L owners and they have objectives and they've been, especially when you look at some of the largest enterprises, it's not uncommon to have someone working with five different BPOs.
NikolaSo I think, what's one more in the shape of an AI vendor.
MichelleYeah. It's a seesaw problem. I think why offshoring came about, why all of these types of things between, operational sort of cost and making sure that you're being responsible, fiscally responsible versus like actually serving a great customer experience. And I think our goal is to get it from seesaw to just flat line, but really high, a really high bar.
NikolaYeah, I think that it's been really interesting. I think we have we have a lot to give there. I
Michelleagree. I agree. Awesome. This has been a great chat. I think our next one, we've got a lot of things to cover. We announced our generative AI platform beta today. And I think that is going to likely be the future, like the evolution of sort of where all of these technologies are coming.
Three years ago, Simplyhealth's PolyAI deployment was only resolving simple FAQs over the phone. Today, they run voice, in-app, email, chat, and WhatsApp conversations on the platform. Nikola Mrkšić talks with Paul...
An agent that gets better every day compounds into a very different agent within months. That's the idea behind Wren, PolyAI's new release: an agent whose job is to make your other agents better. Some of what it finds...
Companies are pouring money into AI, but not everyone is getting the returns. Ian Jacobs, VP and Lead Analyst at Opus Research, thinks he knows why. In this episode, your host Nikola Mrkšić sits down with Ian to work...
When PolyAI built its Builder MCP, one client helped shape it: Vixxo Facility Solutions, one of the largest facilities providers in the US. In this episode, guest host Damian Sasso talks with Vixxo's CTO Derek Neighbors...
PolyAI just launched Dialog-RSN-1, its first audio-native model, and your host Nikola Mrkšić sat down with its builder, Matt Henderson, to unpack why it’s a game-changer for building voice agents. Most voice AI either...