Anne Bibb
Welcome back to unexpected journey. This week we have Liz with Hi operator, Liz went to MIT at the age of fifth teen where she built robots and 3d printers. And after completing her Bachelor of Science in Mechanical Engineering and materials of science at MIT, she went back and got her Ms. at the MIT Media Lab. And Liz worked in physical commodities trading industry in Geneva and Singapore, cuz she’s just an underachiever clearly. And then she went off and founded high operator in 2016, high operators a customer service as a service solution that allows businesses to handle client tickets faster and more accurately through the power of human and AI technology. Welcome, Liz, we’re so happy to have you.

Liz Tsai
Thanks for having me. And super excited to be here today.

Anne Bibb
So excited. And you know, very much because of what’s happening in the world, you’re the perfect guest to have here because there’s a lot of talk right now about AI and automation. And, you know, I wanted to bring you on. Just really specifically, let’s let’s start out with what is the difference between automation and AI?

Liz Tsai
You said a lot smarter when you say AI, don’t you? Yeah. So we really think of it as AI is a subset of automation, right? Automation is kind of the goal is how do you automate things and make them better, faster, more effective, more compliant, and all of that. And AI is really one tool among a 12 of many things that you should have in order to accomplish this end goal of automation, right. And for that reason, we often say automation before AI, right? There’s a lot that you can do with subglottic, simple integrations, simple triggers simple automations. Before you really think about using AI, which is relatively computationally expensive, versus just hooking two things together and telling them to go.

Anne Bibb
That is very helpful, because I think that there’s a lot of confusion out there, especially around AI today. So we’re hearing a lot about, well, let’s just pull it out there, Chad GPT. Right there, the internet has just been abuzz about it for the last few months, people are really mainly falling into one of two categories. They’re either geeking out and truly excited about it. And they just want to play with it, and see what it can do. Or they’re absolutely terrified of it. I’ve seen I saw it actually somebody the other day say that they were losing sleep and throwing up because they were so scared about what this meant for them and what it meant for jobs and what it meant for the future. Where are you on this? What’s your take here?

Liz Tsai
Yeah, so we are squarely in this super excited, you know, innovation should be celebrated. There’s a lot of interesting stuff in there. Now, what I will say is that, you know, there’s the exciting innovation part, I do think that this is maybe a little bit similar to the first wave of 3d printing where everyone went, we can do layer by layer manufacturing, where’s the replicator gun coming in? Right to build a heart? Yeah, is that the we can 3d printers shoot, that’s awesome. You’re not quite at, you know, 3d printing a heart quite yet. Right? In the same sense, where, you know, we have a lot of, you know, lane assist, and, and, you know, enhance cruise control. And there’s a lot of new self driving tools there. But that’s exciting as things you celebrate, it’s going to make humans better and safer and more effective, what they’re doing. But you’re not really anywhere close to that sense of full, general AI quite yet.

Anne Bibb
The way I’ve always thought about it is that AI is really an assistive measure, to helping let’s put it in the context center view, right, like when we are talking to our frontline agents, that we can take some of the easy stuff off of their plates with AI and let it handle the repeatable easy things. So that things that need human empathy and human thought process. They can handle that and let AI handle those easy things and that from an E x and c x perspective, that that will make things easier. What is that where you are on that?

Liz Tsai
100% I think it all comes down to if you take everything a customer service agent does and you break it down, ask yourself know where are the places where software can get you the most leverage, right? Maybe harp on self driving a little bit right? Things like Lane Assist, stay within those highway lanes for 30 miles. Humans aren’t necessarily the best at that because we get distracted and notifications come in, and oh, there’s an interesting fast food restaurant we just drove by. Right? But software is really good at that. And I think in the context center environment, you have a lot of things that are similar like that. Right. And, you know, some examples of this we’ve seen are, you know, things like Grammarly, right. You know, Grammarly is a good one. Yeah, exactly. It can go and read that, you know, six paragraph reply that you’ve drafted up and tell you exactly where you’re making grammatical mistakes, right, and charge up three, in some ways, I think as the next level of that they can read, and whatever you think you want to write, and they can make it look better. But what it can’t do is know all the policies and processes that US agent know, and you know, you want to do, because it’s a very general tool, right? It’s not customized to whatever you’re supporting, or the brand that you’re working on.

Anne Bibb
It also doesn’t know what the other person on the end of the line is going through, you know, it doesn’t know that, you know, your dog is sick, or that you had a crappy day, or that somebody rear ended you on the way home and that you’re in a foul mood, it doesn’t know these things, so we can’t adjust tone to accommodate for them.

Liz Tsai
Yes, US Asian, you need to decide intent, and you decide what’s happening. And then something like chat, GP three, kind of the filter that can make it better, right? Cameras have gotten a lot better, your selfies have gotten a lot better than the Polaroid.

Anne Bibb
Gosh, and the filters, thank goodness, my filters,

Liz Tsai
is that the end of the filter is what perfects it. Right, that takes

Anne Bibb
years off of our lives, right? I mean, that’s what they’re for. That’s what the technology is for. Yeah.

Liz Tsai
And it’s in a customer service perspective, right? You decide as the agent, what you want to do for the customer, you know, enforce the policies, and then charge up three takes a response and makes you an award winning novelist,

Anne Bibb
that, you know, the more that I think about this is that it really is making it so much that it’s going to take such so much of the easy stuff off of the plates. That’s kind of changing the job of what a customer service agent will be in the future.

Liz Tsai
Yeah, taking a time consuming stuff off, right? Well, we the way that we like to think about it is, you know, making a decision on what you do, whether you’re going to, you know, go above and beyond for a customer or whether you’re going to send them something special or make a policy exception, that is a fast decision for a human to make. And it’s also an appropriate decision for human customer service agents who make because they have all that degree of intuition that AI doesn’t have, right. But things like ingesting a ton of text data or eating, you know, well written grammatically correct Well, phrase sentences, that takes humans a relatively long amount of time. But software relatively short amount of time. So that’s where you really want to double down on using AI to make the human better, faster, stronger.

Anne Bibb
So by this logic, though, we are changing the scope of what, you know, theoretically, right? I mean, this is all theoretical right now. But if we’re able to accommodate this with AI, take the easy repeatable things off of a customer service agents plate and let them handle the things that need the human touch, then, by that logic, this should not be a minimum hour wage job any longer because we’ve taken those easy things off. And we’re giving them things that require logic, require human thought process require them to use all of these things in their brain that don’t meet all of these minimum tools any longer. So it shouldn’t wait. Now we’re going to talk about pay policy, right? Like we’re no longer thinking $7.25 cent an hour position for customer service agent.

Liz Tsai
Yeah, because it’s all about leverage, right. I think the reason why tech is so powerful is that it allows people to create leverage, right? Software developers create a lot of leverage in their roles, because every line of code that they write, can be duplicated and does a lot right. I think there’s something to be said about that same model in the Customer Service sense, where customer service right now is a series oriented task we do and interaction to do The next interaction and do the next interaction, right. And there’s, in some sense, a limit to how much leverage you can have in that, and how many, you know, texts or phone calls you can take a day, there just really is physically a limit. You know, when we think about what automation does, and were able to roll that in, we think of the future as you the customer service agent, you’re supervising 10 cyborg agents, right? You’re supervising cyborg agents, automate agents where they can’t do everything, but they’re doing a lot of the work for you. And then you as the human customer service agent, you’re stepping in and putting the finishing touches on determining intent, doing things like that. So you’re not producing one agent’s worth of work. You’re producing 10 agents worth of work with the use of automation and technology.

Anne Bibb
That’s fascinating. As exciting future like it. That’s an exciting thing. How far away? Do you think we are from something like that? I mean, is it tomorrow? Is it like, do you think we’re next year? I think this is like a 2030 thing? I mean, it’s not

Liz Tsai
a step function. It’s an evolution, right? I mean, we do a lot of that, you know, selfishly at high operator, where everything we do is what can we do to make one agent produce more than one agents worth of productivity. And we’re seeing examples today where we can have one agent do you know five agents worth of work. And that’s where it gets exciting, right? Because the manufacturing parallel here is, you know, you, you can have assembly line workers, or you can have mostly fully fully automated assembly lines with one person who’s supervising a lot more work being done than they can do by personally doing all the work, you

Anne Bibb
need to figure out how to get this into homes, where you know, the homemakers can utilize this to be able to duplicate and clone themselves, because that would truly be amazing. There’s just like, if we could figure out how to have one person accommodate five things in the home that like, that would be a miracle right there.

Liz Tsai
Yeah, robotics whole another story,

Anne Bibb
the Jetsons are coming. I want to take a few minutes here to kind of let everybody really get to know Liz, have you ever played the game this or that? Ooh, boy,

Liz Tsai
I’m not great on the spot.

Anne Bibb
So this or that is, it’s a game where I give you two words or two phrases. And you pick one that resonates most with you. And you tell me why. So really easy one to start with? Show or tell?

Liz Tsai
Show? I’m terrified of public speaking.

Anne Bibb
Really, you’re so good at it.

Liz Tsai
If you knew what was happening inside.

Anne Bibb
All right. All right. Now I think I know the answer to the next one as well. But I’m going to ask anyway, book smarts, or streets or street smarts, street smarts. I did not see that coming. Really? that fascinates me. What’s What’s the reasoning behind that one?

Liz Tsai
You know, it’s the same as customer service, where all the technology in the world, and all of the numbers in the world won’t get you anywhere because we live in a people world, right, and start creating the experiences and shaping conversations. And in need the data you need the tools. But at the end of the day, it’s how you interact with other people and the experience that you create.

Anne Bibb
I like whole new level. I mean, I’m already mad respect to you, but hold their level of respect for you right now. Loose guidelines or clear directions,

Liz Tsai
clear directions. Guidelines aren’t invitation to find clear direction

Anne Bibb
and loose guidelines, or to me an invitation for chaos. So I just and I being the OCD person that I am cannot handle chaos. So I am right there with you. So everybody heard it here. We are on the show, the street smarts and the clear directions guidelines here. Okay, back to our questions. So, you know, what do you think is the future of customer experience meeting and merging with AI? Where do you think we’re going here, especially heading into 20? I mean, we’re right now at the very beginning, we’re only two weeks and by the way, happy New Year. How do you think 2023 is going to go with customer experience in AI?

Liz Tsai
2023. I think we’re seeing the beginning of a lot of really exciting potential. Right. And I think we’re going to see this shift around 2023 A theory meets reality. Right? So all this cool potential, you know, you have generative text, you have, you know, a chatbot that you can have humanoid conversations with. The next big challenge is how do we apply this very specifically, very directly? And then, you know, top of mind for everyone, how do we drive business results out of it? Right. I think it’s going to be that shift in Yeah, and two things here. I think one of them is that, you know, Chuck GPT, GPT, right, it’s generative AI, which was was really good at is taking a broad amount of information, learning from that, like, You’re a really smart human who can read really, really quickly. And then learning from that and presenting something out. But if you think about customer service, customer service is very specific, you can be a customer service agent with 20 years of experience. And if you go start working at a new company, you have a whole new set of policies, procedures, and brand voice that you need to go learn. Right? And catch up at while a very impressive tool is not personal and specific and contextual in that way quite yet. So I think a lot of the ShakeOut this next year is going to be people saying, Okay, how do we take that ability to seem like a human, but then combiner with the very specific policies and processes that we need to actually drive customer service results, to create something that we can use that can be customer facing that our our agents can actually use and utilize on a daily basis? I think a lot of it’s going to be that.

Anne Bibb
And with regard to Hi, operator, you talked a lot about and you mentioned cyborgs, which that just, I know that I know, we don’t have cyborgs. But you talked about how you have your agents doing work from basically multiple people? Do you see high Bryter growing substantially over the next year or two? And, you know, what do you see in it coming for high operator?

Liz Tsai
So definitely a lot of growth? Because it’s really the theme that we’re living in, right is how do you deliver great customer experience that no moves all of your NPS and customer LTV metrics and the right direction, right. And you do need humans in the loop for that. But at the same time, how do you drive the business results of that not costing an arm and a leg, right. And there are really only a couple of ways to call it contain your customer service costs, right, you can decrease your contacts, which can be done well it can increase self serve, you know, increase sort of self serve FAQ and chatbots and things like that. You can pay people less, which is never the goal. Or you can make you can pay your agents well, and use all the technology at your disposal to make them much better and faster. And that’s where I think we’re very aligned, where when we say you know, a cyborg agent, right? It’s really saying everything that a customer service agent does. What can we automate? Where can we rely on AI to make them better and faster? So that the overall result is a lot more productivity and leverage?

Anne Bibb
Absolutely. Well, it’s exciting to see where high operator has gotten to so far, just over the last few years, and where it’s going to that person that is listening right now. The one that is on the other end of the spectrum from us that is not geeking out that is terrified, and is worried that AI is going to take their job and is is going to make them you know, just completely irrelevant to everything. What would you say to them?

Liz Tsai
Well, a couple of things. But I think one of them is that charge up three, all of these are models that can only do things that they have learned, right. And you as a human customer service agent, you have very specific experiences and wisdom and ability to discern intent that can’t be deciphered just from reading, you know, millions of pages of web articles, right? That’s essentially what it is, right? It’s essentially saying I’ve ingested a whole lot of information. I don’t have very much context. I don’t really have a North Star. But here’s what it looks like you as a human customer service agent or attorney or whatever you are, you bring that level of experience and discernment that AI really just doesn’t have for you right now. Right. Now, the other thing I would say though, and I’m maybe a little bit surprised that more companies aren’t concerned about this is you know, there’s a lot of if you are building on top of open AI, there’s a lot of vendor lock in there. There really is no additional model out there. And I think we saw this play out over the past couple of years with the Facebook ad See, I have friends data, right? Where there are a lot of apps, businesses built on the ability to use Facebook’s API to grab very specific friends data and things like that. But the moment Facebook started getting negative PR around it, they pulled all of that access. So I do think that for this to really big, but be a big player, we do need to start seeing alternative options that are public. So there isn’t 100% vendor lock in to something like chat GPT.

Anne Bibb
So to put it another way, there needs to be competition. There needs to be somebody else, not just Chappie to GPT that other people can use to build their platforms on as well.

Liz Tsai
Absolutely. Otherwise, I would be terrified. If I was building a business purely on, you know, the desire of one company to continue providing API access.

Anne Bibb
Yeah, I think about there have been a few situations over the last two or three years, especially when the whole world was at home. When one thing went down. One thing I want to say, I’m not gonna name names, one thing went down. And all of the sudden, all social networks were down at, you know, like 16 Different companies went down, because one thing went down. And if everybody’s building off of that one API, I could see that happening as well. And we can’t have that kind of bottleneck going into the future.

Liz Tsai
Yeah, absolutely. I mean, there’s apparently a recent all hands meeting at Google, where, you know, employees asked, Hey, given the popularity of Chad GPT, is Google launching anything similar? Right? And I believe the alphabet CEO actually responded and said, you know, yeah, Google actually has similar capabilities. But we’re not making it public right now. Because if something goes wrong, there’s a ton of reputational risk to Google. Right? What you take that you say, Okay, well, what happens to all of this access with very proprietary models, once we get any whiff of bad PR.

Anne Bibb
So so we’re being driven by fear of what could happen? Intentionally? Yeah. Interesting. It makes me very curious to know how many are actually out there. That simply because we now know that there’s at least one, but there’s definitely more that we just don’t know about, but are potentially afraid to let the world though, for fear of PR issues?

Liz Tsai
Yeah. All right. Is there a competitive edge? Right? Why would you, you know, short of a, you know, great PR, and a viral story? Why would you necessarily uncover your competitive edge?

Anne Bibb
Very interesting. I think it’s going to be a very interesting 2023. And I look forward to having a very similar conversation, either at the end of 2023, or at the beginning of 2024. And kind of mapping to see how it looks. I am curious, though, you, you have a lot of clients that utilize your services for this very reason of trying to be efficient. And do you have any case studies that, you know, without revealing client names, clearly, that you can share with us about how that return on investment to just try and ease some Eastern people’s fears?

Liz Tsai
Absolutely, because what I will say, which is kind of interesting, and I’d be curious if you hear the same is we rarely have clients say, I want to use AI and automation. What we hear instead is that one, I don’t want to miss out on something that can make my business better. But I’m looking for the business results. I want my customer service to be faster, we want better first response times wanted to be more cost effective. And we want it to be higher quality. Right? At the end of the day we are solving for and goals, the tool we use to get there, whether that’s 90% automation and 10% AI or 5050 is what business leaders at least from what I’m hearing that people are looking

Anne Bibb
for 100% They’re not coming to you for the tool. They’re coming to you for the result, and the result is efficient, top level, high quality customer experience.

Liz Tsai
Yeah, exactly. You don’t buy a hammer, you buy the builder, you do whatever you’re trying to do. Right.

Anne Bibb
I bought my house. I really don’t care how they made it. I just want a house With a roof over my head,

Liz Tsai
exactly, I have no interest in knowing exactly how we built that no

Anne Bibb
effect. If they start, I’m probably going to fall asleep because it’s that boring to me.

Liz Tsai
Exactly. So, so those are some of the results that we’ve been able to deliver, but only because of the technology we use, right? So we’ve been able to take programs that are normally with me know, maybe 100, h 100, agents worth 100 seats worth of customer service, right. And then we’re able to use technology, automate bits and pieces of that still lead the human customer service agent in the loop, but cut that down to something that’s, you know, a 20 full time equivalent worth their program. Right. And we do invest a lot more tech in that and engineering time before our clients, you know, we’re able to drive a ton of cost effectiveness, oftentimes No, sometimes cutting their costs in half, almost by doing that, in a way that doesn’t impact customer experience, they’re still getting the human in the loop, and is in many ways higher quality, because you know, who doesn’t make mistakes, calculating things like price adjustments software, installing that human in the loop to deliver the end, empathy and experience in conversation?

Anne Bibb
Can we get a little clarity on something you just said, though, because you said it doesn’t impact customer experience. But I actually think that it does impact customer experience, and it impacts it positively. Even more so than, say, having 100 voice agents or 100. Agents headcount that having those 20 With your technology is actually a more positive customer experience than not impacting customer experience.

Liz Tsai
Yeah, it definitely positively impacts customer service, you’re sure that we think of impacts on residents as negative things. But it definitely positively impacts that. Because part of it is speed, right? You’re happier when you hear back from a company that you’ve reached out to more quickly. But some of it also is accuracy. So a very tangible example is we work with someone complex device with a fairly difficult troubleshooting process, right. And this was a huge manual that customer service agents had to go and learn in order to then troubleshoot this device with the customer. And it used to take a lot easy take, you know, sort of five conversational back and forth to get that customer to your resolution. Because, you know, sometimes the agent would miss one particular part of what was happening, and it would take a couple of back and forth to get there.

Anne Bibb
That honestly sounds like not just a time issue. It sounds like it’s a pain in the ass is what it sounds like.

Liz Tsai
It’s not fun for anyone, right? The customer service agent is following this giant thread of what’s going on and trying to work through the matrix of hey, how do I fix this? The customer is sitting there with a device that doesn’t work, and they just want to get it fixed yet, so

Anne Bibb
they’re not familiar with them. They’re trying to answer the question, which, let’s be honest, they probably answer some incorrectly anyway.

Liz Tsai
Yeah, cuz you’re trying to read through a 20 page. Customer Service agent, right. But software is a whole lot better at mapping and understanding that manual than a human customer service agent, a human customer service agent is. And so that was an example where we took that we boiled it down to our software, automated a lot of the steps of what to check and what you read. And we’ve gotten that down to like two and a half back and forks. Wow, right? It’s a win for everyone. It’s faster for our clients, it’s faster for us. So we’re making more money, and they’re paying less money. And it’s a much more positive experience for the customer, because they’re getting to a working device. And half the time.

Anne Bibb
I’m thinking of myself on the other end of that line. And I would much rather only have two, two and a half exchanges, then five or six. Yeah, because I every single exchange that happens, I get more and more frustrated and anxious and irritated. Yeah. So that is significantly reduced, which means a better customer experience less time, which is more a better customer experience, because every customer equates their time with their money.

Liz Tsai
Absolutely. You’re asking them to invest in your device or hate. And that’s that’s a high ask. But in all of that there’s a ton of automation and some AI behind the scenes and digesting that and creating a better prompting system for the agent. But the agent was still core to that experience, and delivering what it was that a customer needed. And that’s where we see us are the perfect orchestration of letting software do a software upgrade at digesting a whole lot of information. And the agent do what the agent and humans are best at which is providing that end experience for the customer.

Anne Bibb
Fascinating. So, along those lines now that we’ve heard this, what advice would you give to the frontline teams on how to embrace that AI that you were just talking about how to work with it, instead of against it?

Liz Tsai
Remember, humans are great at humans are great at spotting exceptions and providing empathy and allows you say, No, this was a few months ago, but I was on the floor. We believe that every person that Hi operators should have some understanding of what being a customer service agent looks like. So we rotate through the floor on occasion, I was on the floor. And I overheard an agent tell another agent they were working with on a pre automated client. And they said something like, Hey, you need to make sure to check these things. Because that is why we are the human in the loop solver can do a lot of things. But you need to be here to make sure that we’re ultimately serving the customer correctly. software provides policy software provides recommendations, we can’t just click through, you have to be there to provide the human touch, because that is how we generate value.

Anne Bibb
I love that the human I was like,

Liz Tsai
Yes. I love that sounds like you understand why you’re here. That is the role of the human in an automated world is to provide what can’t be automatically surmised from an interaction.

Anne Bibb
I mean, that phrase, the human in the loop, that is phenomenal. Yeah, it needs to be in all your training material. You’re the human in the loop. That’s amazing. You’re critical,

Liz Tsai
right? You gotta make sure your self driving car doesn’t drive itself into the barrier.

Anne Bibb
That’s another show. That’s another show.

Liz Tsai
up so true, good customer service is relatively low reward the high risk, right? You don’t you do absolutely have opportunities to delight and go above and beyond and really solidify the experience in the customers mind. But your average customer who just needs to process the return of our warranty, they just want that done for them. Yeah, right. You’re not split? Yeah, just help me out. I’ve got things to get to, I’m on my lunch break, let’s get this done. But customers get very upset. If they’re asking for, you know, an exchange, and you process to return to a refund. Right now, it’s not anywhere near the same level as driving into your barrier. But it does mean that in customer service, we have to almost go above and beyond to be human and avoid making mistakes, because that’s what the consumer is judging you and the company that you were supporting on?

Anne Bibb
Yeah, well, this, this has been great. And I really appreciate you joining us on unexpected journey of the conversation about AI and automation. And I really do look forward to having it again in a year to see kind of where 2023 Took us. So question on, you know how, first off, why would our guests want to reach out to you? Or why would they reach out to you and to? How would they reach out to you.

Liz Tsai
So reach out if you’re excited about how you accomplish automation and you want to get the end result of automation when it comes to customer service. We yes provide you know, a tech enabled outsource customer service experience that we think about customer service all day every day. And we’re always excited to just chat about that and talk about what you’re seeing and what the correct applications and highest leverage applications are for automation and AI. And you know, you can find me at high operator.com or the company on online or on LinkedIn or just email me at Liz at high operator.com. Always happy to chat. Perfect.

Anne Bibb
And as we have always said we will have links to all of those links in the description of the show. And we look forward to seeing everybody again next week. Thanks for joining us on unexpected journey.

Liz Tsai
Thanks for having me.