Listen and subscribe to our podcast:
In this episode of Customer Service Secrets, Omar Pera, CEO at Reply AI, joins Gabe Larsen to discuss how to use artificial intelligence to make customer service faster and smarter. Omar’s background is in solar engineering. His first job in customer service was at CERN, a physics lab in Switzerland, where he deployed the internal customer service tool for 2000 employees. Later on, Omar moved to New York to work with Big Civil, a tech startup. On nights and weekends Omar worked with his brother, Pablo Pera, developing some indie apps that became really successful with more than 50 million downloads in different markets. Fast forward to 2015, Omar wanted to know the penalty fee for a flight change, and after a terrible customer experience Reply AI was born as an easy way to automate conversations. Omar has extensive knowledge in the technology sphere and provides valuable insights on how technology and artificial intelligence can help create an exceptional customer service experience. Listen to the full podcast episode below.
Instantaneous is the Expectation of the Customer
Every customer service team knows that the customer wants their issues resolved, quickly. Customer experience is becoming more important than price and product when it comes to loyalty. Businesses who prioritize customer experience are the ones succeeding and the ones with more engaged customers. Omar shared a story about being on hold for 45 minutes and waiting 7 days for an email just to find the answer to his question about a flight change fee. This experience motivated him to build a platform that would help automate the most common question businesses receive. Customers want instantaneous solutions to their problems and they want to do it themselves. Omar states, “67% of customers prefer self-service over talking to an agent. … I believe that companies should be focusing on self-serve.” Intelligent automation as part of the customer service strategy not only keeps your customers happier, it also improves the efficiency of your agents.
Chatbots and Other Efficient AI Applications
Chatbots are typically the first platform thought of when it comes to customer service, and for good reason. Almost every business uses charbots to reduce unnecessary engagements with agents. Simple problems and questions can be answered much faster with chatbots. However, Omar points out that there are some other good examples of how to use AI in customer service: agent assistants, automatic categorization of tickets, smart routing or deflection. Omar summarizes:
“Everyone thinks that chatbots can solve everything. Not at all. Really, you can use a chatbot just to gather information and then hand over to an agent. That’s already helping your agents to be more efficient and the customer doesn’t need to wait for that initial “hello” from the agent to start solving their issue. Or, going a little further, you can do some API calls, integrate with your backend, and fully resolve issues”
There are so many different ways to use AI that companies are missing out on because they are only focusing on using chatbots to answer questions. The positive customer experience of the future has AI integrated into several stages of the customer journey.
How to Start Integrating AI into Your Company
Starting the AI journey is no easy task. There are so many things that can be automated, which is intimidating. Omar understands how hard it is and shares, “Start with one, one automation at a time. Focus only on that one… If you have live chat for example, why not understand which topic has the most volume and start gathering information about that today. Maybe a chatbot that answers “Where is my order?”. Ask the customer what is their order number and then hand it over to an agent. Once you control that, move to the next step: integrate with your backend so that later we can give them the tracking code… Start small. Start today.” After doing an audit of your company and the processes that should be automated, pick one, and get started. It will be a gradual process but Omar assures that it will be well worth it in the end. Good customer experience should revolve around making sure the customers get their answers with little friction and in a timely manner.
To learn more about AI and automation in customer service and how to integrate it into your company, check out the Customer Service Secrets podcast episode, and be sure to subscribe for new episodes each Thursday.
Listen Now:
You can also listen and subscribe to our podcast here:
Full Episode Transcript:
Making Customer Service Faster and Smarter With AI with Omar Pera
Intro Voice: (00:04)
You’re listening to the Customer Service Secrets Podcast by Kustomer.
Gabe Larsen: (00:10)
Alright, welcome everybody. We’re excited to get going today. We’re going to be talking about artificial intelligence. We’re going to be talking about some recent news that happened here at Kustomer. We acquired a company called Reply.ai and they bring a wealth of information in the bot deflection and artificial intelligence category. And so we brought on Omar Pera, who’s currently the CEO and founder of Reply.ai and is taking on a Director of Product Position at Kustomer. And we want to dive into this idea of chat bots, AI, and how companies are using these different tools to be more efficient and more effective. So Omar, I appreciate you joining. How are you?
Omar Pera: (00:56)
Thanks for having me.
Gabe Larsen: (00:58)
Yeah, I think this will be exciting and congratulations on the big announcement.
Omar Pera: (01:03)
Thanks so much. This being the most exciting two days of pretty much my life except my wedding, so it’s been pretty good.
Gabe Larsen: (01:11)
I love it. Well, yeah, tell us just real quick. I mean, it’s obviously been a busy couple of weeks. Well, we probably better start a little higher. I wanted to jump right in the announcement, but maybe let’s start with just a little bit about you and Reply. Can you tell us a little bit about what Reply does and a little bit about your background?
Omar Pera: (01:31)
Yeah. So, my background is in solar engineering actually. I started my career in bioinformatics, then moved to CERN, which is a very famous physics lab in Switzerland. And actually my first job in customer service was at CERN. It was a physics lab that I deployed the internal customer service tool for 2000 employees there. And then after that I moved to New York, to work in a tech startup. And then on the nights and weekends, Pablo Pera, which is my brother too and co-founder of Reply, myself and a small team, we created some indie apps that got very successful in the markets. And then after that I jumped, diving into Reply.
Gabe Larsen: (02:15)
I love it. What was the reason to move to New York? Was it just wanting to — I mean, obviously, originally, Spain is home, correct?
Omar Pera: (02:24)
Well it is home today. I’ve been around eight years outside Spain. I went to first UK, then Germany, then Switzerland for CERN, and then I moved to New York–
Gabe Larsen: (02:38)
Oh yeah because CERN would’ve been in Switzerland. Yeah that’s right
Omar Pera: (02:38)
And then I moved to New York, to be working with Big Civil, which is a company that got acquired for a bunch of dollars. It was a pretty exciting moment there. Then later that’s when I started Reply.
Gabe Larsen: (02:58)
Okay so that makes more sense. So you’ve bounced there, so you’re quite the international traveler, it sounds like.
Omar Pera: (03:03)
I am. But I still have a funny accent. I cannot deal with that.
Gabe Larsen: (03:08)
It just depends on where you are if the accent is funny or not. You know, I lived a couple of years in the Middle East, a couple of years in Germany and my accent was weird to them there as well. So, no harm, no foul. It sounded like you did when you mentioned you had successful Indie apps, it was very successful. I mean to the tune of millions of downloads, is that correct?
Omar Pera: (03:29)
Yeah. We deployed around 10 or 12 apps to the markets, on many different verticals. It was just our own ideas and some of them actually got into the order of millions to a total of 50 million downloads.
Gabe Larsen: (03:44)
Oh wow. Wow. Hopefully you were charging a couple bucks per download on those.
Omar Pera: (03:47)
No, I know. I wish. I wish, but no, it was a combination of advertising and other things.
Gabe Larsen: (03:55)
If it was, we probably wouldn’t be talking right now. Right. If you’re doing five dollar– So I’m glad you didn’t. I’m glad you didn’t do that. Well let’s get into Reply then. I mean you kind of walked us through the background and then you ended saying, we then started Reply. So what was the story behind that? How did it start? Why Reply?
Omar Pera: (04:16)
Yeah, so it was funny because one of the apps, [inaudible]. So from five emails, we got into 50 emails and then 100 emails per day. So we could not keep up. We actually hired on a freelancer or other services there, some contractors, but really we could not keep that momentum. And then actually one day, this was 2015 one day I wanted to know the price of the penalty for a flight change. So I called an airline, and waited on hold for 45 minutes and then the phone disconnected and it was super annoying. And then I sent an email and waited for seven days to respond. So I really knew that that was a bad experience.
Gabe Larsen: (05:01)
It was seven days, huh?
Omar Pera: (05:03)
Seven days to respond just to understand it was 70 bucks of the penalty fee. So we really knew that that was a very bad customer experience. We also knew that from being a customer. So we hate to wait, like everyone hates to wait. So we had a look at the market and said I don’t think there is a platform out there today that makes it very easy to automate conversations. So that was the beginning of Reply. We now are a customer service automation platform now integrating to Kustomer. We help companies to resolve their most common questions without any agent in the region.
Gabe Larsen: (05:42)
Oh my goodness. So important. I just feel like — it’s funny because I’ve often debated this kind of bot versus, deflection bot versus human interaction. That’s been a conversation for a while. But man, in the last couple of months, the amount of energy and the amount of focus that companies are putting on doing more with less kind of to use your line to eliminate that agent. It just seems like it got — it was important, don’t get me wrong, but now it’s like, wow, it’s really important. Right? It’s just kind of the times have changed. So, one more followup to that. I think you’ve kind of hit on this, but I want to just double click it and that is, what then really was the core thing you were trying to solve for Reply?
Omar Pera: (06:35)
Yeah. So we experienced both sides of the problem, right? As a company, as business owners and as myself as a customer. So these gave us unique insights, how to fix the problem because first, for any company of the planet who is growing, it’s very, very hard to scale customer service. That’s pretty much a fact. So we sit down to set ourselves to solve that problem. And the ultimate goal is pretty much to have a great customer experience and make customer service faster and smarter. That’s what we really want to do. So we started doing that on Reply by providing a do it yourself tabled platform. And then we quickly realized that there was more than just chat. So we’ve expanded our offering to provide self-serve on all channels, including chat, email and your contact form. And I think to date, which I believe is going to be around 1 million and a half consumers have gotten any answer from us without waiting.
Gabe Larsen: (07:41)
Wow. Wow. Okay. So those are the number of answers or deflections or self service inquiries that have been solved.
Omar Pera: (07:47)
Yeah, the number of customers who have been touched with our automation and being helped.
Gabe Larsen: (07:53)
So over a million is where you’re currently standing. And the great thing is it isn’t just, that last point you mentioned, it isn’t just the bot, it’s actually email you said and what was the last one?
Omar Pera: (08:05)
So we can replace the contact form by using your FAQ. So the FAQ is usually the most creative piece of self-serve that any company has. So we try to make use of it so that whenever you are submitting a contact form, we provide you with suggestions for that contact form. So they don’t need to wait.
Gabe Larsen: (08:25)
Fascinating. And again, just so timely and maybe I’ll follow that up with that. I mean, why do you think it is so timely? Why is it so important now that people start to think about these types of tools, these types of technologies?
Omar Pera: (08:41)
Well, COVID is one of the causes, which is very sad, but at the same time, it has caused that customer service is totally overwhelmed. And the reality is that people still expect that you solve the issues fast.
Gabe Larsen: (08:57)
Yeah.
Omar Pera: (08:57)
They are used to one day delivery. So maybe now with a three day delivery, they’re going to find that not with a one week delivery. Right. So I will say that today for even any business consumer company, if you have a contact form or an email and you have three to seven days response time, you know that you have a problem. Right? So I think it is — every company has learned to do social very well, but usually customer service has been a call center right in the background of every company. So I think now more than ever, customer service is also part of your brand and the only, not the only, there are many solutions out there, but, AI and better self serve can really help get your company in a very good position to have a good customer experience.
Gabe Larsen: (09:48)
Hmm. Interesting. I mean, it does seem like there was some good — and maybe you have those off the top of your head, but there’s some good stats out there that really highlight the need for this new age world, which is people do want answers more quickly. They expect it. To your point, it’s a little more part of your brand. You have a couple of those off the top of your head.
Omar Pera: (10:11)
I have one which is the most funny one is that consumers would rather go to the dentist than contacting a company by phone. That is hilarious. And then you mix that with I think it’s 67% customers prefer self-serve over talking to an agent. So I think you put those two together and you have everything that every company is struggling to maintain that level of productivity. I believe that companies should be focusing on self-serve.
Gabe Larsen: (10:47)
Wow. Yeah. That idea that 67% of consumers would prefer to do self service. Why do you think that? I mean, truthfully, I believe that. I’m a consumer, we’re all consumers ourselves. And I think I have my own reasons. But what was the why behind that? Just in your own opinion, why do you think people want that?
Omar Pera: (11:09)
So I believe today we are very impatient and we want answers as fast as possible. Even the rise of messaging. Everyone is texting, is doing texts these days, right? With WhatsApp or Telegram, all those messaging apps that are popping up, they have even more traffic than social today. So I think that instantaneous communication, you want that translated into the brands that you love.
Gabe Larsen: (11:43)
Yeah, yeah. I mean we all — that instant, I love that word. It’s instant gratification. Sometimes it’s used in a negative context, but truthfully, I want it, I want it now and some of the current environment I don’t think has helped that. So let’s see if we can get into a couple examples here. I love the idea — I mean, you’ve got a strong background in artificial intelligence, machine learning. AI is definitely a big talk track in customer service today. Obviously it fits into some of the things we’ve talked about doing more with less. What are a few examples of using AI in customer service? Let’s see if we can click down there.
Omar Pera: (12:21)
Yeah. So today when you say AI in customer service, everyone is thinking about chatbots correct? But I think we should not end there and maybe we should not even start there. So we will have that. But they believe that there are many, many good examples of AI including chatbots that you can apply to your company. So one of them could be like agent assistant, right? You have an assistant who is helping your agents to be more efficient by suggesting answers on real time and it can even learn from past behaviors. So it’s really helping you out. Another one, very simple is to categorize your tickets automatically. Like how many companies do have one or two people categorizing those tickets in order to get to the right person or even later for insights. The most typical one today who has been a big one in the phone space has been smart routing, who to direct this issue to based on what you know. That is a big one.
Omar Pera: (13:33)
So we’ve already gone through three and then the fourth is usually chat bots. Correct. Everyone thinks of chatbots as a chatbot that can solve everything. Not at all. Really, you can use a chatbot just to gather information and then hand over to an agent. That’s already helping your agents to be more efficient and the customer actually to not need to wait for that initial hello from the agent or you can actually go a little bit further and do some API calls, integrate with your backend and fully resolve issues. I think the last piece, which I believe is the one where on Reply, we’ve been focusing the latest, which is deflection with your FAQ. Everyone has an FAQ today and you can really provide recommended suggested articles in a very smart way and try to really provide that paragraph that really works and resonates with the customer asking a question so that they can get resolved that issue faster.
Gabe Larsen: (14:32)
Got it. Okay. So we got agent assists, we’ve got, categorization of conversations, routing, chatbots and some of the deflection I think is what you talked about. Couple of follow ups on that. One is, that is a lot. So if you were going to start, if you were going to recommend for someone to start somewhere on their AI journey, which one of those initiatives would you maybe start them on?
Omar Pera: (14:55)
So I will say that usually the biggest problem is that you have too much volume, right? So the simplest way to set up something is usually deflection because you already have an FAQ, right? So that’s usually the first step that I will suggest. And then if we move into — if you have live chat, that usually depends on the channels. If you have live chat, why not understand which is your one topic that you have more volume on and start gathering information for that topic today. Just a little chat bot that says, Hey, maybe it is “Where is my order?” questions are your top one. So let’s just do a chatbot that’s only that. Ask the customer what is their order number and then hand it over to an agent. You can do that in a day, today. Later, okay, let’s integrate with your backend so that later we can give them the tracking code. Okay, good. Next step, we can get the tracking code, we can call the shipping provider and we can get them the exact day that it’s going to be delivered. So it’s an iterative process and there is no risk. So I think that’s important. But the biggest piece of advice that I usually give to every single company is to start very small and start today.
Gabe Larsen: (16:19)
I liked that iterative process, right? Because I don’t know if there is a way you can just — I think truthfully that’s maybe one of the mistakes I did when I first was playing around. I’ll use the chat bot just as an example. I thought, you know what, I can just throw it up on my website and be done. That was a mistake. You kind of needed to go through what you were talking about; build, measure, repeat. I learned that the hard way, but I think that’s a great principle to live by. Let’s see if we can continue down that journey. So as we think about getting AI into CX, I love your iterative process, but you were also saying, Gabe, as we’ve worked with companies, I like this idea of starting small and starting today. I want to continue down that journey. What are some, what do you mean by that?
Omar Pera: (17:03)
So this means that for someone, one of our customers, you saw a large marketplace and they have many, many different types of issues. From, where’s my order to scheduling issues, to refunds, change orders. So there is a lot. So what we need is just to start with one, start with one, one automation at a time. Focus on that one. Don’t focus on any other topic. You can focus on two or three. Okay. But it is even better to just do one and then — but the most important part is if you do one you do that very well and everything else reduces that friction to the customer. So that’s very important. For example, when I mean they could start today and start small; do you have an FAQ contact form? Okay. Start today with an FAQ deflection. What are you waiting for? You can remove even 40% of your questions with that. Or do you have live chat?
Gabe Larsen: (17:58)
40% of your questions? Oh, so they wouldn’t put them in the form, they could just get answers via the deflection.
Omar Pera: (18:04)
Suggested articles. Yeah. If you have a good knowledge base, you can get suggested articles and usually customers are not browsing anymore a lot of knowledge spaces because usually they are tired of searching and not finding anything. Right. So if you bring AI to bring the answer to the question at the right moment, they are going to pay attention.
Gabe Larsen: (18:24)
Hmm. Got it. Got it. Okay. So start small. Start today. I love some of those examples, whether it’s contact form or live chat, you can get something that resolves your order or deflect something. Okay. Where do you go next? What’s kind of — after you start small, start today, what’s your next principle to get people up and going here?
Omar Pera: (18:40)
So the concept I would say is pretty simple. Make it very easy to find answers. Make it very easy. That means that you have a knowledge base, open it, very visible, not in the footer.
Omar Pera: (18:55)
Their knowledge base. They write all these great articles. No one ever uses it, right?
Omar Pera: (18:59)
Correct. And they have the contact us page on one side and their knowledge base on the other like put your contact page, inside your knowledge base. Make your articles short and sweet. Simplify it. Cutting to short sentences. This is not marketing, this is more about really going to the point to solve that question and I think it’s very important. These are common mistakes where I see like the knowledge base is a structure from the point of view of the company. I think flip that. Flip the navigation of your knowledge base into the point of view of the customer and the common problems that they are going to have.
Gabe Larsen: (19:33)
Yes. Yeah, it is interesting. It does feel like — you’ve mentioned a few other things but it does seem like often the knowledge base is hidden. It’s kind of back in the middle of nowhere. Sometimes I have a hard time figuring out where the knowledge base is and the thing that I love that you guys do or that you’ve talked to me a little about is just bringing that knowledge base to the customer. That may mean some of the deflection capabilities, but you know now I don’t have to go search. I may be chatting with a bot and the bot is actually bringing that knowledge base from the back to the front to me rather than having me go find it. I love that idea. I just think that’s so easy. It’s to your point, it’s getting answers easy and reducing friction. Okay, so we’ve got one, two, three. What is, what’s the next one? Where do you go next when you coach people on this journey?
Omar Pera: (20:17)
So I believe this one could also be one of the most important parts if you don’t want to damage your brand, which is don’t frustrate your customers. This is such a simple idea. And everyone usually will say that they are not trying, but if you do any automation, you need to be not careful. You need to have a good process. If you do any automation today, like dude, never pretend that you’re a human ever.
Gabe Larsen: (20:45)
Oh, interesting.
Omar Pera: (20:46)
Yeah, that’s, that’s my point of view. There are many others who believe in a different way, but I have a strong point of view on that. Another strong point of view which I’ve also seen in the market quite a lot is the, “Sorry, I don’t understand” message right? Chatbots have been a little bit controversial in the past few years going in ups and downs because I believe it is not about the technology was not there, not like if you improve it has improved a lot, the technology. But I think the process is very important. So the process to create a nice conversation design brings that maybe you don’t need that “Sorry, I don’t understand” message. Forget that. Get out of the way. Automation, get out of the way very fast. Go to an agent directly. If you don’t understand a question, reduce the friction. So I think that is very important. And then just to put a concrete example, remember the chatbot about “Where is my order?” that I mentioned before? If the customer says absolutely anything that is not related to “Where is my order?” you go to an agent directly without waiting, without doing anything, without rephrasing, just go to an agent. Maybe ask their name and email and go for it right away.
Gabe Larsen: (22:03)
Wow. Yeah. So you are, don’t mess with the bot, get to the agent faster. Don’t mess with it so much is what I’m hearing you say there. Wow. All right, well, last question before I let you go and then I would love to hear maybe a quick summary. But, again, congratulations on the news. I’ve never had a company acquired. I’m sure you’re kind of on cloud nine. Can you give us, from your perspective, it’s been a busy couple of weeks the last two weeks, but how did it all kind of go down and, are you excited about the result? Are you excited about the next step and joining Kustomer? Give us kind of your quick take on the last couple of weeks here.
Omar Pera: (22:49)
Yeah, so I actually told my brother last year in a Slack message that I just saw a demo with Kustomer and I was so amazed, so amazed about the product that I believe that they will pretty much eat the whole market share of digital customer service with a powerful CRM. So, I’m actually — I think most interpreters usually say that you are going to feel kind of sad in a way. It feels sad. But I think I have more excitement about the future. So I believe now it’s like I’m in this new adventure. It’s like starting on day one with the same excitement as if I started Reply on day one. So I am very, very, very happy of this.
Gabe Larsen: (23:39)
Good man. Well again, congratulations. I know everybody here at the Kustomer Crew is excited to have you join. So talked about a lot; the acquisition, best practices with using AI, a lot about bots and different AI capabilities. Summarize kind of the conversation for us as we think about customer service leaders trying to win and take today’s challenging times.
Omar Pera: (24:01)
So I would say any customer service leader knows that it’s very hard to keep satisfaction high while you are growing. That is very hard. So I would say make self-serve a priority in your company. Start very small, make it very easy to find answers and that, you bring that together with a customer service tool that is very good. Put in the context of a contact in the same screen such as Kustomer. I believe it’s actually the key to success. So my last point will be, really the technology is there. There’s no need for one year plans. You just need to start a small start today and really don’t wait, just implement these broadly and the benefits will come over time.
Gabe Larsen: (24:45)
Yeah. Yeah. By small and simple things great things come to pass, a wise man once told me that. So Omar, really appreciate your time. If someone wants to get in touch with you or learn a little bit more about some of the things you’re doing, what’s the best way to do that? Like LinkedIn maybe or?
Omar Pera: (25:00)
Well omar@kustomer.com, omar@reply.ai, Twitter. I am Omar Pera on LinkedIn. Any time, I have DMS open on Twitter. So I really liked talking to people. So yeah, look me up on any channel.
Gabe Larsen: (25:20)
All right, well it sounds like you got a couple options there. So Omar, thanks so much for joining and for the audience. Have a fantastic day.
Exit Voice: (25:35)
Thank you for listening. Make sure you subscribe to hear more customer service secrets.
