in

Creating a better CX | Etie Hertz from Loris

If you’ve needed support through a live chat, you’ve probably wondered how much of the person on the other end is actually a person. If it’s not a full-on bot you’re talking to, it’s probably someone pushing out prepackaged responses.

The results can often be flat and frustrating. Etie Hertz didn’t initially set out to solve this problem, but a nonprofit that helped improve crisis hotlines had inadvertently stumbled on a technology solution the private sector badly wanted.

Loris was born—a tech startup that pairs machine learning and empathetic responses to help guide support engines toward more satisfying interactions and better results.

More information: https://loris.ai/


Etie Hertz has 20 years of hands-on experience building, advising & scaling Technology-led organizations. 

He is the CEO of Loris.ai, Inc., which  empowers the world’s fastest-growing companies to have better conversations with their customers. Loris is an AI assistant for text-based conversations (chat/SMS/email) that plugs into existing customer service platforms (Zendesk, Salesforce, etc.).

Its real-time coaching software equips customer service teams with live guidance, helping agents respond to requests more efficiently and effectively, ultimately improving CSAT and customer LTV. Loris customers have seen a 25%+ boost in productivity, a 33% improvement in quality, all while ensuring consistency across agent performance and brand voice.

He was previously the Founder & CEO of Payment Revolution which provided payments technology and support to over 7,000 businesses and processed over $2 billion of transactions annually. Payment Revolution was acquired by ShopKeep, Inc., and Etie became the Chief Revenue Officer at ShopKeep, driving revenue growth from $15 million to over $40 million annually, serving almost 30,000 businesses. ShopKeep was acquired by Lightspeed POS, Inc. (NYSE: LSPD).

Prior to that, Etie was the co-founder and CIO of the Ronen Real Estate Fund, a real estate development fund focused on multi-family developments in and around the tri-state area. Etie began his career as an M&A and Venture Capital attorney, and practiced at two large, multinational firms.

Etie holds an MBA from Columbia University, where he graduated with Dean’s Honors, a JD from Benjamin N. Cardozo School of Law, where he served as the Supervisor Editor of the Law Review, and a BS in Finance from the University of Illinois at Urbana Champaign.

DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!

Etie Hertz 0:00
Because what happens when you do typical training, you’ll, you’ll sit there a couple hours and then you’ll forget it within three or four days. So the only substitute for that really is a constant reminder of what to do what not to do so that that’s how we pivoted to to a real time platform.

Alexander Ferguson 0:19
Welcome to UpTech Report. This is our apply tech series UpTech Report is sponsored by TeraLeap. Learn how to leverage the power of video at teraleap.io. Today, I’m joined by my guests, Etie Hertz, who’s based in New York. He’s the CEO at Loris, welcome, it’s good to have you on. Thanks so much for having me, Alexander. Loris is a real time coaching technology that sits on top of a customer service platform. Help me understand ej, what was the genesis of this idea of a real time coaching technology for customer service platforms? Like Where did this come into existence? Yeah, so

Etie Hertz 0:51
we actually started a nonprofit in in the City, New York City called Crisis Text Line, which is largest crisis hotline, it’s SMS based start about seven years ago, backed by some pretty influential people around the world doing really amazing work. And all SMS based huge database, hundreds of millions of messages. And a few years ago, some private businesses started reaching out, saying essentially, like, you know, we’re moving heavily to digital channels from phone based, and you are the best in the world at talking to people in written form in text, SMS email chat, maybe you could teach us how to do that. So that was kind of what spawned the idea to spin this out into a for profit business.

Alexander Ferguson 1:29
So it started with basically consumers or individuals helping them as a crisis line. And you’ve built up a wealth of knowledge of what are the best ways to interact with with individuals, and then on the public side, or the business side, they’re like, wow, we need to we want that insight. So you’ve How did you then extract that and provide that as a real time coaching technology.

Etie Hertz 1:50
So the first actually iteration of the idea was to build training videos, empathetic training videos for large enterprise clients. But then through testing with some huge brands, folks, were saying, Listen, you have this huge database data scientists, if you could do this live in real time. Everybody would need this, you’d have to, you’d be negligent not to use this, right? Because what happens when you do typical training, you’ll, you’ll sit there a couple hours, and then you’ll forget it within three or four days. So the only substitute for that really is a constant reminder of what to do what not to do. So that that’s how we pivoted to real time platform.

Alexander Ferguson 2:23
Learning as All right, we let’s just turn it into videos, people watch it, and then they learn and then you realize No, no, no, they don’t learn they actually need it all the time, consistent right there with them. So you turn that into that technology. Just give me a quick visual of how does that in play? How does it look? How does it work?

Etie Hertz 2:39
So a big part of our philosophy is we don’t want to change the way that agents are used to working. So you know, a lot of companies say come off your platform come by this new integration, this new platform, switch all your systems, it’s pretty hard to do that in enterprise. So from from the beginning, we’ve been pretty steadfast about sitting on top of folks existing platforms. So we integrate the systems like Zendesk and Salesforce and my person in those in those types of platforms. So that the experience for the agent is essentially the same, they sit and they see the same screen, but we have a layer that sits on top of that screen. So in real time, we’re suggesting techniques that we think they should be saying language we think they should be using, and in real time, we’re giving them feedback of how it’s going.

Alexander Ferguson 3:18
based off of whatever that conversation is how it’s happening, they’ll see a little indicator, oh, you should probably say this next, or they may be feeling this. So you should respond this way.

Etie Hertz 3:29
Exactly. Exactly. So what’s happening now is, well, a few different things are happening at the same time. This shift to digital we thought would take years, it’s taking young COVID took a few months. And to add to that as folks are shifting into digital in a big way. And currency is becoming a thing. So agents are expected no longer to do one ticket at a time, but three or four or five at a time. Wow. Imagine this in a stressful time period and agents around the world trying to speak in the same voice, it becomes really difficult to do. Full stop, right? And then you have to think about how do I do that? Well, and how do I remember what I said to Xander? How did what I said to Stacey, which ticket in my hand? What issue is this? right if there’s a huge cognitive load that you have to undertake, so it’s pretty cool.

Alexander Ferguson 4:12
So the trends here are that there’s definitely an increase in demand an assumption from from customers like I want real time support from a customer remotely through text. So now, companies are need to scale their customer service team without getting more people ideally, and but to make sure that there’s a consistent consistency across all of them. And that’s where your technology comes into play that it allows to have that consistency and help them respond faster and more effectively.

Etie Hertz 4:44
Right. So oftentimes, we think that the agents job is a difficult one to begin with. Now you’ve taken someone who spent many years learning the phone and they’re now typing and then maybe their English is not their first language. This is not an easy task to begin with. Now handle a few of these at the same time and You know, people are more stressed out, customers demand everything now. So now you’re juggling and you’re trying to go fast and trying to stay on brand and it’s it oftentimes breaks down.

Alexander Ferguson 5:10
That’s a great point, people’s expectation is definitely real time support, like, give it to me. Now, if you haven’t responded within a few minutes, what’s going on, even if anything, damages the brand,

Etie Hertz 5:21
right, exactly. So if I come, you know, people want live, consumers want live, right. And yet, the stats would say 79% of consumers want to have synchronous conversation with their brand. Maybe 10% of companies are ready for that. So you as user coming to a brand new you expect live now I’m on chat, right? I should have a live experience. And now you’re waiting for a long time. And it’s slow. And it’s not what you expected, you get upset. So now you got an email and phone, and then you created three tickets at one issue. So it’s a spiral, right? And it becomes an exponentially difficult problem to solve.

Alexander Ferguson 5:51
It only gets gets worse from there. Can you can you share any of your customers, just like how you’re helping any any particular ones?

Etie Hertz 5:59
I mean, so we work with fast growing brands like like freshly lift, and folks in the spaces that are onboarding a lot of agents quickly around the world. And struggling to keep up with the demand. Right? They they’re at the forefront of technology, they know they need real time. They’re just scaling really quickly.

Alexander Ferguson 6:16
Now this this whole concept, this ability to coach, that’s a funny word to say a computer coaching a human. I mean, we understand actually training computers using people like hey, this is correct, this is wrong, but having a computer tell someone else this is right. This is wrong. How How have you been able to extract that learning from that that Crisis Text Line? Can you share anything about what the technology behind that and how it applies? This?

Etie Hertz 6:42
I mean, so the initial learnings were essentially like, how do you speak to people in conversation? That’s the general, if you think about a building, that’s the foundation of our building? That’s the first floor in general, how do you empathize? How do you validate? How do you put yourself in someone else’s shoes because people want to be heard, and oftentimes, when you’re juggling too many tickets, you lose that right? 10 minutes into a shift, you’re getting yelled at by three different people, you can’t even remember where you are. So that’s kind of the base. And then in the last few years, I’ve been working with these really a fast growing companies on huge data sets, learning a lot of information about specific use cases, what things work, what don’t and our coaching, doesn’t tell you, Alexander, we think you should go left it says we should you should go left, potentially, or maybe right, or maybe you should speed up or slow down, and then the the agent will agree or disagree with us. So we’re getting smarter.

Alexander Ferguson 7:25
So if anything, it’s giving a couple options for the human to make that final decision. You know what actually, I feel like this option right here makes the most sense out of what I see.

Etie Hertz 7:34
Exactly. So we want to think about it like, it’s like, Agent assist. It’s like an autopilot, right? If you’re sitting in a self driving car yourself to touch buttons, and you still have to grab the wheel once in a while. But the experience should be easier for you. My analogy, actually, when somebody says the self driving cars, I see we actually have you drive three self driving cars at the same time. Right.

Alexander Ferguson 7:52
At the same time, now, I feel like I heard something about like a real time score as well, like how your conversation is going, can you can you explain that?

Etie Hertz 8:01
Sure. So we score not just a message, a lot of companies, you know, you get a C set NPS or QA gives you a score oftentimes, after the fact. So it’s not actionable in real time, most companies would get that maybe eight to 15% of the time. So we score not only 100% of messages, but 100% of turns within a conversation. So in real time as an agent, you will see Alexander was upset now he’s doing better the conversation is going well keep going doing great job. So a little encouragement, you know, folks don’t realize that these agents, even if they’re sitting in a humungous room, maybe in a BP other walled off from one another, they don’t really interact, it’s a pretty difficult job you’re doing almost by yourself when it’s live. So getting a little bit of encouragement, you know, taking some of that load off really makes an impact.

Alexander Ferguson 8:45
I remember the exact word used in the technology phase, is it semantics? Or is it the ability to understand the human emotion in amongst words? Is that what you’re able to pick up on?

Etie Hertz 8:58
Yes. So, again, you know, you have to think that when you’re doing so many tickets, and maybe you’re not interacting in your native language, really understanding what you’re asking me takes some time and it can be frustrating for both sides.

Alexander Ferguson 9:10
What about with people like sarcastic? Mo, like nine All right, thanks. That was good. orgasm is

Etie Hertz 9:15
an issue in NLP in general. Yeah. Especially in different countries, even English speaking countries, right. There’s different sarcasm in the UK versus the US, Australia. So that gets Yeah, that can get interesting. But it’s it’s, you know, small use cases. But yeah,

Alexander Ferguson 9:29
it’s like constantly iterating upon that, but it’s that’s what comes back to the end user to be able to use it is taking a step back then of kind of the bigger perception of of you. People are coming for support, and they’re they’re expecting this and particularly one industry I think of maybe could use it more like FinTech. There’s, I think something in the story around Robin Hood last year, like just one individual who he couldn’t get support and then just That type of situation, do you think having more Customer service is like the key is like just being able to get more people to more accessible? If you have questions, what are your thoughts there?

Etie Hertz 10:11
It’s a very, I mean, it’s a broad? It’s a very, very deep question, right? So a lot of CX is thought of as a cost center. You know, if you’re a manager of CX, maybe I just want to deflect, I’m gonna send you a knowledge base, you have a question, here’s a link to an answer. But again, as we said, more and more end users expect resolution, I’m invested in his brand. When I signed up, I didn’t speak to a salesperson, I downloaded an app. So my only interaction as a user really is with the agent, right? That’s, that’s my brand experience. So you know, if I’m going to spend my time trying to solve something, I expect you to be here and help me. So that that’s, in my view, kind of the baseline of where we are in the customer service world today. In FinTech, specifically to your question, people have this perception that FinTech is just a bunch of rich people trading money, but in reality, it’s really you know, all these technologies are unlocking services for folks who are underbanked underserved. Millennials, people don’t have necessarily money. So if you get stuck, your account is locked, you know, you can’t access your your theory, that could be your life savings. And if you can, you can imagine how stressful that must be. You don’t want to get a knowledgebase article sent to you. Right, you want to talk to someone, you want them to listen to you, and you want them to respond. So you know, the Robin Hood gets really sad case of that going completely wrong. They this customer couldn’t find anyone to talk to them on making sure there was a phone for them. There was a phone number, I’m not sure there was anyone on the other end. But you know, the x it’s basically goes down to the expectation brands are expected to do certain things today. Right? consumers are expected to sign up seamlessly start spending money, start doing all these things brands are expected to kind of be there when things break down.

Alexander Ferguson 11:45
That’s a fascinating way to put it is, as a consumer, we’re expected sign up right now start using it. So the same expectation should be in return as a consumer. Alright, brand be right there with me holding my hand right now. Right? Not someday later, like I would sign up one day for your service. Right? Exactly. I also appreciate your point of this, this the customer service person, maybe the first immediate connection as as part of that brand. What do you can you can you see going forward in the future of customer service? The how it may change with technology? And what should one expect, especially people who are in charge of the customer service team in a company.

Etie Hertz 12:30
So I mean, I think specifically the language, the conversation between the agent and that end user is going to be the type of language that is the most interesting for everyone. Because again, they’re very few sales conversations happening today. You don’t go to your local store and sign up, right. So those those interactions are so interesting, they tell the brand everything they should know about their customers, their product, is anything that working, how do people actually feel about you. So I think marketing is gonna get involved in that cx leader will become more important. And all the data that emanates from that pipe essentially, is going to be the most valuable data that exists.

Alexander Ferguson 13:03
For you guys, specifically, is there any particular size of company or type of company that you are focused on serving right now? Like, I’m just wondering, anybody who’s listening to this? And like, is this for me? Is this a good solution for me? How would you identify?

Etie Hertz 13:17
So we typically say, you know, fast growing companies that have 50 plus agents and rolling? If you’re, if you’re abroad, or expanding in different places, you’ll feel the effects immediately? Also? There’s certain I mean, yeah, that’s basically I mean, that’s kind of kind of broad. We used to say, you know, if you’re having difficult conversations, if people are working remotely, but those things are all pre COVID. Now, those things are all they apply to everyone. So

Alexander Ferguson 13:42
you have to everyone, so it’s really 50 customer agents is kind of where that’s where this type of solution comes into play. Makes sense? Exactly. As far as then the integrations you mentioned earlier have two main integrations right?

Etie Hertz 13:55
We have four that are completed, Zendesk, Salesforce, Twilio flex in my person are working on several others. So those cover a pretty large swath of the customer service population at the moment.

Alexander Ferguson 14:07
And it used the initial training set that you were able to build this entire, quote, real time coaching off of came from the the text line support line. But if I understood correctly, you’re actually trying to train it. Also now with with your customers to continue to get better. Am I hearing that correctly?

Etie Hertz 14:25
So we are, we’re building all kinds of different models based on all kinds of different things. And yeah, as we get more and more data sets, and we’re working with a lot of universities, we’re getting smarter and smarter. Language is changing, issues are changing. So as we get smarter, the output should get better obviously,

Alexander Ferguson 14:42
you can share our what you’re excited about on your roadmap, what’s coming up new features and where you’re headed.

Etie Hertz 14:49
Wow, so where so many things. So we’ve been talking just about real time, but how do we empower real time so today you know, until recently, you Didn’t have any way to impact conversations live as an organization. So you’d get some insights, maybe do some training, that would be the end of it. So if we want to empower you to do that live, how do we do that? So we’re we’ve actually about to release content management systems. So heads of content teams can come in and suggest language to us that can go through our system, we can suggest that to agents, we can see if that works, score and tell you that work, that sentence is better than this sentence. You know, so you can actually in chat, do things that were impossible before, right? People came from, at least in written form from email, and an email, you have macros. So customer care teams write these long macros, and they send them to you. And they hope that they land, right. But oftentimes, it sounds like you’re tone deaf as an end user, because you’re just getting the same snippets back. In live chat, you’re expecting something else, it’s dynamic. You can’t just send me these like wrote, you know, eight sentences in a row. So how do you how do you ensure that it’s dynamic that the end user doesn’t feel like he’s talking to a robot. So our system ensures that and we take input from management so that they can suggest language, we put it through our models, we score it, we give them insights to actually go back and put it into the system. So in real time, we just keep getting smarter and smarter. So it’s kind of a flywheel that continues to get better.

Alexander Ferguson 16:12
So basically, the content teams can write could be able to write content of for different answers. And then your system can say, Okay, this is good content will probably fit here. And then the actual agent, when they’re in play, this content will just suddenly appear as people are asking questions, they Oh, this content may actually be relevant to what they’re asking right now.

Etie Hertz 16:29
Exactly. And then the agent will agree or disagree, and then the customer will agree or disagree, so Well, no, right? We’re gonna get smarter and smarter about what the end user how the end user feels.

Alexander Ferguson 16:39
So your system also learns, if the agent keeps not using that option, that it’s for some reason, it’s not a good option. Yes. Wow. So it’s actually multiple learning data points. They’re both conscustomer. And the agent all taking context for then the content writers to be able to

Etie Hertz 16:54
exactly what happens, what happens now, if you think about agents that communicate synchronously, they’re told certain policies, but 90 plus percent of what they write is the free forming. And they go off brand, they’re saying whatever they want, oftentimes, if you’re like, in a BPO, someone gave you a nice sentence, they claim that that works. And now the whole BPO is using that sentence, it was never approved by anyone, right? So the thing about brands that are growing really quickly, doubling every year, how do they stay on message on brand, keep their policies consistent, as they’re growing, you know, to different countries. And scaling really fast, that’s a big problem folks have. So we found that we’re able to really increase consistency, by doing

Alexander Ferguson 17:33
having back to, to actually just did the the actual use of the product, how it works, these options appear, and they can choose the different options, and able to as the agent, customizing any of those three options before it gets submitted, making suggested management has to approve it, got it, got it. So it’s really it’s this guided experience it that still that they can choose which option is best. And then if one of them doesn’t fit, or that needs to adjust, they can write something and then it goes to August,

Etie Hertz 18:01
you have your clickable text, and that is gonna sit there based on the models and the models are predicting, right, that’s impacted by what we learn what we see what management wants to add, to make sure that the brand stays on message separately, the agent can type whatever they want, they can edit anything that we suggest,

Alexander Ferguson 18:15
yes. Okay, it’s just adding new content to the models. That’s what the manager would approve or disapprove. As fast as fascinating, but the probably one of the nice things because everything’s real time that that analytics, that data is right there all the time, you’re not waiting for your NP MPs or C SAT. Yeah. Or I’m curious, what do you see that coming back, and just this future of of, of customer service is I guess everything is going to go real time, like you’re going to know how your customer what your customers think about you all the time. This is an accurate representation at that moment. Exactly. So

Etie Hertz 18:55
I mean, you know, my experiences, cx leaders, oftentimes they don’t get enough credibility, they don’t get enough praise for what they do. Oftentimes, you know, we’ll say something seems to be broken. Have you heard anything, anyone yelling at you about something? So now you have to go sift through all this data and figure it out. But what if you had all of the information that all of your customers are passing to you and you can actually digest it and understand it simply, suddenly, you can feed information to product into marketing and say, in real time, this is working this new offering, it’s not working? What I’m seeing in the data. You know, you spoke mostly about the real time aspect for the agent. But we’re also building tools to alert management and team leads and QA something’s happening here. Something bad is happening on this conversation you might want to jump in, create a way for them to communicate for brands communicate within the team to figure out a resolution for conversation hasn’t ended yet.

Alexander Ferguson 19:48
If you could make a prediction of technical attack prediction of what what what could we see in five, maybe even 10 years from now, in this space, what comes to your mind

Etie Hertz 20:01
Oh, wow, I can go so many ways. Look, in general, the end users experience is the focus, it’s going to end up getting better and better. You’re going to have mapping of personas, you’re going to know that Alexander’s this kind of person, he started this type of agent, that conversation should take two minutes, and this is probably how it’s going to go. It’s you know, the stuff you see in science fiction movies, it always ends up happening. So it’s gonna be, you know, expectations will keep going up, and he’s not going to get you know, less critical, they’re gonna get more and more critical. And companies are going to have to step up to the plate and meet that demand. Because you see what happens. companies that have something special, that have great software have great products, they grow so fast, the demands on their time, their resources grow exponentially. So they have to figure this out fast. And it’s gonna drive a lot of incredible innovation.

Alexander Ferguson 20:49
It’s a powerful future that you predict. And I agree like what we see in sci fi and science fiction, you’re like, well, that will come to reality because people expect that they they assume that it’s going to come and technology will bring it to reality. Thank you so much, Etie. For those that want to hear a bit more though about Etie’s journey. This is not the first company that he’s led stick around. for part two, we’re going to hear more about his insights of both growing this company and the previous companies. On our second part of the interview. Thanks again for your time today. For those that want to learn more about Loris. So go over to Loris.ai and you’ll be able to look like just try Laura. So I’m assuming be able to get a demo. Get started as a good first step. That’s great stuff. Yes. Awesome. I will see you all on the next episode of UpTech Report. Have you seen a company using AI machine learning or other technology to transform the way we live work and do business? Go to UpTech report.com. And let us know

Digitizing the Automotive Industry | Arjun Patel from WorkClout

A Better Way to Manage Teams | Adam Berke from WorkPatterns