Leveraging AI to Power Your Contact Center With Aarde Cosseboom and Vikas Bhambri

Leveraging AI to Power Your Contact Center With Aarde Cosseboom and Vikas Bhambri TW 2

Listen and subscribe to our podcast:

In this episode of Customer Service Secrets, Gabe is joined by Aarde Cosseboom and Vikas Bhambri to discuss how to use AI in contact centers. Aarde is the Senior Director of Technology and Product for Global Member Services at TechStyle. He’s spent the last decade working in e-commerce and is the author of the book Enable Better Service. Vikas, a familiar guest on the show, is the SVP of Sales and CX at Kustomer and a 20-year CRM / contact center veteran. Both Aarde and Vikas have extensive knowledge on the use of AI in customer service and they have come together to discuss how other businesses can optimize with the help of AI.

“Omnibot”, The Omnichannel Bot

Customer expectations have changed significantly over the last few months, and companies are starting to feel the strain— especially in regards to their AI. While autobots have a reputation for dehumanizing companies, we are starting to rely on them heavily as customer needs increase. To ensure chatbots have a positive impact, Vikas and Aarde focus on making sure they are used as an omnichannel tool. Aarde states, “You can’t just have a chatbot on your website anymore, and it only be in your chat profile. It’s gotta be across all of the different channels that you use to support your members.” As customers switch channels, the bot needs to be available to support your customer on their preferred channel. Gabe, Vikas, and Aarde called this adaptable bot an “omnibot.”

Knowing the need for effective AI, and bots that function on multiple channels, Vikas and Aarde discuss who should build the bots and how they should be built. Because coding and creating AI can be taxing, they recommend finding a good partner to help, as it will be a better use of resources. As for how an omnibot should be built, Vikas notes the need for authenticity to the brand. He states, “If you’re a fun hip brand, you want to keep it relative to that. If you’re maybe a more mature brand, you want to keep it in tune with your … general reputation and what your customers expect of you.” In other words, make sure that the bot matches your brand. And, as an additional note, let customers know they’re talking to a bot. Customers don’t like to question whether they’re talking to a person or not.

How to Humanize a Customer’s AI Experience

One of the main concerns with using chatbots, even ones that are authentically built to the brand, is that consumers lose the human touch of customer service. This is a valid point, but Vikas and Aarde explain ways to overcome that while still increasing efficiency. To humanize a bot experience, have a good team behind it. In regards to AI Vikas states, “You still need people that will go and optimize the program behind it.” It is a team effort to optimize a chatbot, and constant evaluative measures will ensure that it grows and changes with the needs of the customer. Good AI is not meant to replace people in customer service, but to aid those committed to helping customers. In fact, Aarde mentions optimization tactics that fix AI and help the customer at the same time. He says, “When we feed the transcripts to our agents, our agents are actually reading through and seeing where things fail and then they escalate that to the bot architects, the engineers in the background. So they could change those bugs.”

Best Practices and Final Advice on How to Optimize AI

Transcribing bot conversations and having the bots follow the customer across multiple channels helps with the overall customer experience. Additionally, not being hesitant to transfer someone to a live agent is a good tactic. If people are saying “Operator”, pressing zero, or yelling, don’t use the bot to fix the problem, have a person step in and do their job. Aarde’s final piece of advice, or best practice, is to not tackle the hardest type of AI first. Don’t try for voice AI from the beginning. “I recommend trying,” he states, “but trying it slowly. So testing with maybe a low volume channel first, just doing a small portion, maybe 10% of volume, see its success rate and then roll it out to the greater population.” Add AI to your company’s customer service department one step at a time. Agreeing with Aarde, Vikas adds, “Look at your FAQ. What are the articles that people most often go to that resolve their issue?” He also suggests, “[Talk] to your agents or even [look] at the analytics in your CRM ticketing tool to look at, ‘What are the macros they most often use?’” While investing in AI can be an intimidating venture, bots can provide increased efficiency to your company, and successful self-service to your customers.

To learn more about how to leverage AI in your customer service department, check out the Customer Service Secrets podcast episode below, and be sure to subscribe for new episodes each Thursday.

 

Listen Now:

Listen to “Leveraging AI Automation and Self-service to Power Your Contact Center | Aarde Cosseboom and Vikas Bhambri” on Spreaker.

You can also listen and subscribe to our podcast here:

Full Episode Transcript:

Leveraging AI to Power Your Contact Center With Aarde Cosseboom and Vikas Bhambri

Intro Voice : (00:04)
You’re listening to the Customer Service Secrets Podcast by Kustomer.

Gabe Larsen: (00:11)
All right. Welcome everybody to today’s broadcast. We’re excited to get going here. We’re going to be talking about one of these really relevant and interesting conversations, leveraging AI and self-service to really power your contact center. To do that we brought on two special guests. We’ll let them introduce themselves. Aarde, why don’t we start with you?

Aarde Cosseboom: (00:31)
Sure. Thanks again Gabe and Vikas for having me and Kustomer, of course, for hosting. I’m Aarde Cosseboom. I’m the Senior Director of Technology and Product for GMS, which is Global Member Services for a company called TechStyle. And we’re an e-commerce retail company.

Gabe Larsen: (00:47)
Awesome. Vikas, over to you.

Vikas Bhambri: (00:49)
Vikas Bhambri, SVP Sales and CX here at Kustomer, 20 years CRM Contact Center Lifer, looking forward to the conversation with Aarde and Gabe.

Gabe Larsen: (00:57)
Yeah, this is exciting. And you know, myself, I run growth over here at Kustomer. So let’s get in and let’s talk about this. Aarde, let’s start with the big picture. What do AI and self-service bots even solve?

Aarde Cosseboom: (01:11)
Yeah, this is a great question and really hard to answer specifics because every business is slightly different, but I’ll try to stay as high level as possible. Really it helps with self service, it’s in the title, but deflection, reducing contact. There’s a lot of automation that happens as well, too. So not only automating for your customer, but also automating a lot of the agent processes like creation of tickets and then auto dispositions as well too. And then one of the things that’s kind of hidden that most people don’t think about, and it’s actually one of the things that we don’t really measure that well in the industry in this area, is customer experience as well, too. So as millennials and gen X are expecting these types of tools, it creates a better experience for those people who are expecting it.

Gabe Larsen: (02:01)
Vikas, maybe you can add onto that. I mean, why do you think this is such an important conversation more so now than it was even just a couple months ago? Give us kind of that thought process.

Vikas Bhambri: (02:11)
Sure. I think what we’re running into right now is folks like Aarde are really seeing a tremendous surge of inquiries into their contact center. And the reason they’re seeing that is there’s the heightened level of anxiety and expectation for consumers. Most of what they’re shopping for, they want now and it doesn’t matter what it is. In fact, I was talking to a friend of mine who’s in the middle of buying a bike. Now, normally you buy a bike and you’re good. Whenever it shows up, it shows up. But because of the quarantine, he is literally like, “I need a bike so that I can have something to do with my kids.” So when he placed an order for the bike and wasn’t immediately notified when his bike was going to be available, he got extremely concerned and started pinging the bike shop. So I think it’s really interesting to see that behavior, particularly in these times, the ticket surge and putting pressure on people like Aarde and his peers to be able to respond.

Gabe Larsen: (03:20)
It feels like, again, there’s just more need for it than ever before. How do you think about chatbots versus social versus some of these other channels? Do you feel like they’re just different times to use them, is it different companies, is it different industries? Aarde, what’s your thought on kind of the mix of channels that are out there, why people would use one versus the other, et cetera?

Aarde Cosseboom: (03:42)
Yeah. And it goes back to expectations. So your customers expect a lot from you. And as we grow in channels in the customer service realm, growing the social and then direct social, which is things like WhatsApp and Apple business chat, direct SMS, and MMS. Those are all areas that we need to grow into and when we do grow into, we need to create an omnichannel experience. So you can’t just have a chatbot on your website anymore, and it only be in your chat profile. It’s gotta be across all of the different channels that you use to support your members. And as a member switches, as they do the channel switch, maybe they start in chat online and then they say, “You know what, I’m going to pause the conversation. And now I’m going to go to Facebook messenger.” You need to follow that with your AI so they don’t have to start all over from scratch with that automation tool.

Gabe Larsen: (04:36)
I like that. Vikas, how would you add to that?

Vikas Bhambri: (04:38)
I think Aarde nailed it. The term chatbot is so yesterday, right? Your bot needs to be omnichannel, your bot needs to be available, not just via chat as a channel, but you know Aarde mentioned Facebook messenger, WhatsApp, SMS email, right? So when we think about automation and bots here at Kustomer, we think about it regardless of channel, I mean, even email, right? Why is it that somebody sends an email and somebody actually has to enter a response? Why wouldn’t you send some responses that will allow that customer to self service, even by email, which is obviously one of the older, more mature channels. So that’s how we think about bots here at Kustomer.

Gabe Larsen: (05:23)
Well, look, I’m as guilty as anybody; the chatbot I’m so used to thinking chatbot and it’s something on the website. Is there a different term? Is there, I mean, obviously as you guys kind of pointed out, it’s better to think about it, maybe in an omnichannel approach, but Aarde, I’m looking for you on this one, man. How come you haven’t invented a term that is an omnichannel chatbot? What is that term, what is it?

Aarde Cosseboom: (05:49)
I haven’t invented it, but it is out there. It’s IVA which stands for Intelligent Virtual Assistant and really it’s the omnichannel bot experience, doesn’t matter how you use it, but that’s how you deliver it. So Virtual Assistant or Intelligent Virtual Assistant,

Vikas Bhambri: (06:07)
Gabe, I’m not the marketer on this call, but I’m going to give you a lay up here and you can give me credit. And if our friends at Zendesk are listening, they’ll probably copy it as they always do, but Omnibot.

Gabe Larsen: (06:19)
Omnibot! Oh my goodness! Oh, stolen.

Aarde Cosseboom: (06:22)
I like it.

Vikas Bhambri: (06:22)
I’m a transformers kid. I grew up, I’m a transformers generation. So that just sounds super cool to me.

Gabe Larsen: (06:29)
Honestly that sounds like —

Aarde Cosseboom: (06:31)
[laughing]

Gabe Larsen: (06:31)
Omnibot does sound like one of those transformers. What’s the main transformer? What’s the old guy?

Vikas Bhambri: (06:36)
Optimus Prime.

Gabe Larsen: (06:38)
Optimus Prime. Optimus Prime, meet Omnibot.

Aarde Cosseboom: (06:43)
That’s a great name for a bot too. We could brand it.

Gabe Larsen: (06:47)
It totally works. That probably is good for this question you guys. I consider myself a programmer. I wanted to build my own bot. My kids are doing little things with programming. It seems like a lot of people are building bots these days. Should someone just build a bot? Should you buy a bot? And excuse me, an Interactive Virtual Assistant. Aarde, let’s start with you man. You’re out there in the market, talking to people, can companies just build these things? Is that easy or should you buy it? I’m confused.

Aarde Cosseboom: (07:19)
Yep. Great question. There’s a lot of controversy here and lots of different companies are doing their own little flavor. As technology grows and changes, it’s enabling companies to be able to build their own. Things like Amazon Lex or Google dialogue flow, it’s getting a lot easier than it was a year ago or even five years ago. But in the current market and we assess this here at TechStyle every six months, we recommend to buy or partner, is what we like to call it, partner with an actual partner that has the technology in place. You get a couple benefits from it, ease of use, and you’ll get to market faster. You won’t have to do that long implementation, have to have those developers and experts build something from scratch. You’ll be able to lean on the expertise of your partner to help you with that. And then the other thing that’s really beneficial that most people don’t think about is, when you’re partnering with a technology partner, they’re going to be leveraging all of the AI and machine learning that they have across all of their other customers and bring all of that to you and your bot. So if there’s a best practice in your space, we’re in retail, for example, and we use a partner and they have a best practice for another retail customer, they’re going to knock on our door and give us that easy flow without us having to do all the legwork. So I recommend buy for now and partner with a dedicated partner that has it in that ecosystem.

Gabe Larsen: (08:45)
Yeah. Look, it’s becoming, I mean, there’s just, there’s enough out there. You guys, I think you can get it for a good enough price that I don’t know if you need to dedicate a whole engineering team to kind of build your own automation, roles and bots, and things like that. So I don’t think I’d disagree with Aarde. Vikas, this one just came through on LinkedIn, this is from Keith, this question, and I meant to throw this in here and so I want to throw it in now. He said, “Hey, look, we’re trying to humanize our bots. So we designed them to help people not be viewed as an application. But it still comes — begs the question of how do you think about these bots? I’m thinking more on the website at the moment. Do you name it the bot, do you put a human there? Do you — how do you balance that? Have you seen best practices on that?

Vikas Bhambri: (09:26)
Yeah, the first thing that I recommend to customers is you got to keep it authentic to your brand.

Gabe Larsen: (09:32)
Okay.

Vikas Bhambri: (09:33)
That’s number one. If you’re a fun hip brand, you want to keep it relative to that. If you’re maybe a more mature brand, you want to keep it in tune with your just general reputation and what your customers expect of you. The other thing is, I think in the early days, and most companies have gone away from this, I remember there was a brand in the UK that had announced a bot, but they branded it Lucy. Ask Lucy. And customers cannot really tell whether they were speaking to a human being or a bot. And they actually got very negative feedback because people were just asking questions and the bot at that time, you can imagine almost seven, eight years ago, wasn’t trained. It couldn’t answer half their questions. So I think the more that you let your customer know, “Look, you’re dealing with a bot” and that allows them to give some flexibility and some leeway to you to understand that look at some point, this bot may not be able to answer my question; to know that you can always escalate to a live human agent, right? So you can still give it a name, right? But making sure it’s authentic to what it is. And if the point comes where it can not resolve the customer’s inquiry, that they know there’s a handoff, a seamless transition. That’s another thing a lot of people get wrong. Right? So now I connect to the human agent, don’t make me ask the five, six, seven questions that I just went through with the bot. The agent should pick up the conversation fluidly from where I left off. Aarde what do you think?

Gabe Larsen: (11:09)
Yeah Aarde, I want to talk — do you agree because I think you might disagree?

Aarde Cosseboom: (11:15)
No, I do agree. There’s a little bit of uncanny Valley; gotta be careful about not tricking your customer into thinking they’re talking to a human. So I totally agree that you have to upfront tell them that it’s a bot. I like to brand it as giving it kind of a bot accent. So if it’s a voice bot giving it a little bit of a mechanical accent, so they know that it’s a bot or, not having a hundred percent of a fluid conversation fragmented a little bit more so they know that they’re talking. Also, you could declare it at the beginning of a chat or social conversation saying that “You’re engaging with an AI tool at this time.” And then, another key point here is you’re right, try to do it on brand. So we have 95% of our customers are females. So we have a female voice. If you’re selling golf clubs online, you may want a male voice because there may be a higher percentage of males that are listening to or engaging with your bot. So think about voice, tone, accent, especially accents, U.S. accents. So if you’re on the East Coast, don’t put words in there like “cool” or “hip” or things like that. Make sure that it’s localized to your customers and brands.

Gabe Larsen: (12:29)
Yeah, don’t use one of those weird Utah accents like you hear coming in all, all “Here y’all.”

Vikas Bhambri: (12:36)
One other thing to Keith’s question, right? And this whole concept of an application; look, it goes back to back in the day and chat, we started out with what we called a pre chat survey, which was literally, “Here are the five questions you need to answer so that we know who to route you to, who you are,” et cetera. Then it became a bit more where people were doing authentication. And so they had some data. Then we moved to this concept of conversational form, which was still a bot, but it asked the question in a humanized way. So it wasn’t just “Fill out these five questions.” It would ask you the question one at a time and maybe there was a variability where if you said you were a buyer versus a seller, the next question would change. Now Keith, where we want to take it is the bot can gather so much data about the customer before they even type in one word. So a lot of that is now picking up with the information that is now unknown to you so that you can then either answer the inquiry or then route it to the agent. So it should necessarily have that kind of predetermined, almost process flow. You can be much more mature about how you even go about using natural language processing for people to just key in things and it doesn’t have to be hard coded, right? So I think there’s a lot that you can do there now.

Gabe Larsen: (14:00)
I like that. This is, I think, one of the questions that comes up often, this is such a cool feature look at this. I can just throw this in here, right here. Look at that. Are you guys seeing that?

Aarde Cosseboom: (14:11)
Yeah.

Vikas Bhambri: (14:11)
Yep.

Gabe Larsen: (14:12)
Geez louise, man, look at this technology. Scott Mark, little shout out to Scott Mark. What are best practices around the handoff from a bot so we stop dropping the ball? I think that’s — we wanted to get actually into some best practices. Maybe we start it now. That’s just a big debate. It’s when you handoff, how do you hand off, how many questions do you ask? It’s just, it never feels right. Thoughts? Aarde let’s start with you on that one.

Aarde Cosseboom: (14:38)
Yeah, absolutely. And you have to think of one thing first, which we call the IVR prison or the chatbot prison. You’ve got to allow people to get out of that prison. So if you get the same question twice and it’s not — you can’t recognize the right answer like, “What is your email address?” and can’t recognize, ask again, can’t recognize, fail it out to a live agent. That’s a good best practice. Also if they say the word operator or press the zero key on their phone, or if they start cursing, definitely fail them out of the IVR. Don’t keep them in prison. Always allow them a way out of that IVR. But then when you go over into the agent experience and that handoff, even for the experiences where someone engaged with the bot for a very long time, and there’s a long transcript, maybe there is actions that were done like they updated their credit card information with the bot, they updated their billing information, their name, profile; all of that you want to transfer to an agent, screen pop not only the member profile, start to fill out the case or tickets so the agent doesn’t have to do it. And then also, feed them the transcripts so that if the customer or member says, “Hey, I talked to the bot, it updated my billing address, but I think it didn’t do it right. It didn’t do the right street address, the right number. Can you go back and check and see if it did that?” The agent should be able to scroll up through that transcript and see exactly where it failed and then fix that, that failure.

Gabe Larsen: (16:11)
Yeah. Vikas, what would you add to that?

Vikas Bhambri: (16:13)
I think the biggest, so Aarde nailed it, right? So, your initial implementation, those are all the best practices. I think the challenge for most brands is you’ve got to treat this like a program management, just like a marketer would if they were doing a promotion on their website or doing a campaign. Constantly revisiting and optimizing, right? So one, your bot is going to get smarter if you’re investing in the right technology. But two, if you’re finding that customers are constantly getting challenged, that process in your step, go and see what do you need to do to modify it, to smooth that out, right? So where are people cursing, where are people hitting zero? Where are people saying, “Get me to a live human agent?” How do we further optimize that piece before we do it? So I think that’s the biggest thing I see is where people will roll these things out and then forget about them and then six months later, they’ll say, “You know what, this isn’t working and we just have to pull it off the site.” And that to me —

Gabe Larsen: (17:16)
Why do you have to call me out like that? Why do you have to call me out like that? I mean, geez louise. In all truthfulness, that was my first experience with a bot. I mean, it’s been a few years back, but I don’t know. I thought you could throw it on the website and it would maybe like, I don’t know, do its things, some sort of magic or something. And three months later, I’m like, “This thing’s a piece of garbage.” I totally, I mean, I came to the heart of the conclusion that like anything else, it has to be iterative and optimized. I love that one.

Vikas Bhambri: (17:45)
No, I think Gabe, this is an interesting thing, right? Because people keep talking about AI just on a broad macro level. And you know, people will say, look, “AI is going to put everybody out of a job. We won’t need salespeople. We won’t need marketers. We won’t need customer service people.” No, because the role will change because the technology is great, but you still need people that will go and optimize the program behind it. Right? So I think, I think that’s an interesting nuance just as we think about AI generally.

Aarde Cosseboom: (18:11)
Yeah. And talking a little bit about supervised learning; so when we feed the transcripts to our agents, our agents are actually reading through and seeing where things fail and then they escalate that to the bot architects, the engineers in the background. So they could change those bugs. So your team members, your agents are now a part of a QA or quality assurance process on your technology, which is huge. And it kinda levels up the agent as well, too. They’re no longer just answering chats and emails and phone calls. They’re now, they now feel a part of the organization because they have a higher role in reporting this information back.

Gabe Larsen: (18:49)
I’ve been hearing more about this kind of bot, almost like a role, like a bot architect. I love the idea of getting the frontline people in front of it. Guys, give me a couple other nuggets. I think that’s where people want to go with this because I think people are getting onto the idea that they need to have these assistants or bots on their sites, et cetera. I don’t know if people know some of the best practices, lessons learned from deployment, where they get started. Our time’s a little bit short, but give us a quick rundown. Aarde let’s start with you then Vikas, we will go back.

Aarde Cosseboom: (19:18)
Yeah, absolutely. I’ll make it super short, but, it’s a huge chasm to cross from having nothing to having something. That’s why I recommend trying, but trying it slowly. So testing with maybe a low volume channel first, just doing a small portion, maybe 10% of volume, see its success rate and then roll it out to the greater population. So try to do the easier channels first. So online web chat is probably the easiest or a social chat or an SMS bot. Don’t tackle voice first. That’s going to be your hardest heaviest lift and you’re going to be sidetracked.

Gabe Larsen: (19:54)
Vikas what do you think man?

Vikas Bhambri: (19:54)
Yeah, I agree with Aarde. Look, you have to look at this as a crawl, walk, run, right? If you try to bite off more than you can chew, you’re going to end up pretty miserable. So for me, number one is, look at your FAQ. What are the articles that people most often go to that resolve their issue? Maybe that’s something you want to be more proactive serving up. The second is talking to your agents or even looking at the analytics in your CRM ticketing tool to look at what are the macros they most often use, right? Because if somebody is just cutting and pasting, we’re hitting hashtag time after time, again, that means those are probably some, that’s some low hanging fruit that you could front end via a bot, the omnibot, for them to resolve themselves. So those are some things that you could look at. Query the data you have, and then just think about, “How do you want to be proactive and thoughtful about putting some of these things in front of your customers?”

Gabe Larsen: (20:54)
I think that’s spot on you guys. I mean, my biggest takeaway from today, I’m going to trademark Omnibot. That’s what I’m doing. That’s — I could barely listen to you guys. I was thinking so much about money I’m going to be making on Omnibot here. No, I’m teasing. Aarde, really appreciate you joining. Vikas, as always, great to have you on. For the audience, hope you guys have a fantastic day.

Vikas Bhambri: (21:19)
Have a great weekend.

Aarde Cosseboom: (21:20)
Thanks everyone.

Exit Voice: (21:27)
Thank you for listening. Make sure you’re subscribed to hear more customer service secrets.

Why Data Will Power the Future of the Contact Center

Why Data Will Power the Future of the Contact Center TW

Until now, the omnichannel, cloud-based, 360-degree customer view-enabled contact center was mostly a pipe dream, touted by technology vendors and thought leaders, with a majority of businesses falling short of this gold standard. Most customers still expect to fight their way through a dead-end IVR, endure multiple transfers, and repeat their information to agents who have zero context on who they are or why they’re calling.

As technology grows more robust, however, more and more businesses are starting to overcome these bottlenecks, more of which are related to a lack of data transparency. Businesses are using AI and machine learning-enabled platforms to unify their data across the organization, route customers based not only on queues but context, and design self-service platforms that facilitate end-to-end support.

Treat Every Customer Touchpoint as a Potential Data Source

For many businesses, their website is the seat of personalization. By collecting data on customer’s viewing history and purchasing habits, they can provide personalized recommendations and proactive support based on context, such as offering help through web chat to a customer who’s having trouble completing an online purchase. But a truly omnichannel experience is one where personalization follows the customer, whether they’re on the phone with an agent, shopping online or visiting in-store.

This means that data you collect from your website must be reconciled with the customer’s activity in all other channels to build a complete 360-degree view of each individual customer. When an agent interacts with a customer, regardless of channel, they should be able to see the customer’s buying history, sentiment and previous interactions (across every channel), status of their orders and customer’s preferred channel.

Says Kustomer CEO Brad Birnbaum, “Imagine having a conversation with a friend but not being able to remember anything about that friend, or any interactions you’ve had with them previously. It would be difficult to have a truly personal or meaningful conversation. That’s how traditional retailers have historically interacted with their customers, with a large blind spot around customer preferences and history.”

Optimize Human to AI Interactions

“Agents for complex issues, AI for simple ones” is an oft-repeated principle for successful human-AI interactions in the contact center. However, customers still find themselves calling when a chatbot does not function as anticipated. For this reason and others, the contact center is often still considered a cost center rather than a revenue driver. Once businesses learn how to optimize their self-service channels, while giving customers recourse to contact a live agent if needed, agents will automatically become the go-to touchpoint for complex issues and expert recommendations, and thereby come to be perceived as subject matter experts.

Without the burden of responding to repetitive inquiries, agents can focus on building a relationship with the customer. As Birnbaum says, “It will become the customer service agent’s job to reflect the company’s mission and values, and act as a trusted partner. The changing expectations of consumers means that customers want to do business with companies they believe in, feeling as though they are a part of the brand. Customer service agents can help do just that, through both proactive and reactive support.

To learn more about common blockers contact centers must overcome to power the future of customer service, download the CCW Special Report on the Customer Contact Vision for 2025.

Download CCW Report


A Look Back at Customer Service During the 2019 Holiday Season

The holiday season has only just wrapped, and it was arguably the most important season for any retail customer service organization. When there are issues with holiday orders, you encounter the very real possibility of unhappy customers who are angry and embarrassed by missing or incorrect gifts. Customer expectations are at an all time high, and organization must over-deliver during their greatest times of need. Whether it’s handling simple product questions in real time, proactively alerting customers when there is an issue with their order, or rectifying any subsequent issues upon delivery, customer service during the holiday season could be the difference between a lifelong customer and one lost to the competition.

How Peak Season Customer Service Has Changed

The digital and direct-to-consumer shifts have had a huge impact on retail customer service. While customers now expect instantaneous service on a multitude of channels, they also expect personal and helpful interactions, similar to face-to-face interactions with in-store associates. In fact, according to a recent Kustomer survey, 75% of consumers aged 25-34 said they expected personalized communications from retailers. And that is 15% higher than those 65 and older, meaning personalization is becoming increasingly expected with younger generations.

2019 Holiday Season Developments

Last year there was a drastic increase in multichannel shopping (both online and in-store), and multichannel support inquiries are also on the rise. That means consumers may start an inquiry on one channel, and finish it on another. Whether the channel switch is because they are on the go, or they didn’t get a prompt response on their original channel, multichannel inquiries can cause duplicative work resulting in agent collision as well as the unfortunate need for customers to repeat information. It is important that retailers have a strong omnichannel support solution in place for any peak shopping period. A true omnichannel support solution can integrate your combination of communication channels in order to capture the free flow of conversations across channels and display the data in a single screen. This ensures seamless transitions and consistent experiences from one channel to the next.

Peak Season Challenges

The constant struggle for customer service organizations during peak shopping periods is sheer volume. With more shopping comes more support inquiries, and businesses that don’t have a scalable strategy in place, supported by the right technology, may not be able to deliver on customer expectations. Additionally, many businesses hire “seasonal employees” to help with the busy periods that they must heavily train to ensure they provide a consistent brand experience. With software that does the heavy lifting for them, providing unified customer history in a single screen and delivering standardized responses via dynamic content, the onboarding burden during an already busy time will be lessened.

Additionally, with a high volume of inquiries, customer service organizations often have trouble prioritizing the most urgent or pressing issues, and simply stick everyone in a queue, which is often unbearably long. Retailers can use AI and automation to intelligently route the most pressing issues to the most appropriate agents, or even prioritize loyal customers.

Overwhelmed customer service organizations often fall into the unfortunate habit of delivering bare minimum support in order to complete inquiries as quickly as possible. It’s important to realize that during peak shopping seasons, your customers are also stressed out, and expect retailers to deliver on their usually stellar service just as thoroughly as they would on the slowest day of the year. That means delivering real-time support, on any channel they choose, in a personalized manner.

Overcoming Peak Season Challenges

Whether it’s Cyber Monday, Valentine’s Day or Back to School season, brands should not only have a scalable strategy in place, but also technology that enables them to be more efficient and effective. AI and automation can improve the precision and speed of service by automating repetitive, manual tasks. While there is always fear of losing personalization when using AI and automation, with the right platform, businesses can actually do the opposite. For instance, if a business leverages customer data properly, AI could ask personalized questions based on an individual’s purchase or browsing history. These interventions save time for both the customer and agent, and increase the time spent on the actual issue rather than information gathering and low-level support.

To learn more about the results from the 2019 holiday shopping season, and how to properly prepare for future peak shopping periods, download the full report.

Download the Report


What Does the Future Hold for Contact Centers?

It’s undeniable that customer experience is only becoming more central to business success. But being thought of as an “icon” in the customer service space is challenging. A surefire way to succeed? Prepare for the future now. According to a CCW Digital survey of contact center professionals, the future of customer service is already within grasp. Read on to learn what pros in the industry think, and how to prepare for the future.

Channels of the Future

Just fifteen years ago, the iPhone didn’t exist. Neither did Twitter. In five years, the landscape of customer service channels may be dramatically different. According to the CCW study, the phone will not disappear as a dominant support channel, with only 17% of respondents expecting its relevance to decline, however a digital transformation may come to fruition.

A whopping 84% of respondents believe chat and messaging bots will become more central to support functions in the next five years, showing the importance of AI to the future of customer service. But instant agent channels will also become more popular, with 81% expecting messaging to rise and 76% saying live chat will be imperative. Lastly, social media and connected devices will have a place in the future, with 68% of respondents saying social will become more important, and 60% saying connected devices are on the rise.

The Customer Service Intelligence Challenge

Did you know only 11% of consumers believe that organizations take their feedback seriously? It’s true, and it’s an issue that businesses are trying to solve.

Collecting great insights from customers is the leading contact center objective that companies want to achieve by 2025. Unfortunately, only 1% of organizations currently believe that their customer intelligence strategy is perfect. Why? Almost half (47%) of organizations have data scattered across various systems, showcasing the importance of a unified data environment to gather meaningful and actionable insights.

Other customer intelligence issues reported by businesses were: not collecting enough data (43%), not using data to personalize the experience (43%) and not doing enough to understand customer sentiment (40%). An additional 40% believe that they don’t even HAVE sufficient data to fully understand their customers.

The contact center of the future will have to leverage systems that unify data and give businesses a seamless way to analyze customer intelligence and take action on that intelligence.

Establishing Customer Service Objectives

Many of the “objectives of the future” are the ones we strive to currently measure, but may have difficulty achieving. In order of importance, customer service experts think the following will be the biggest objectives by 2025: reducing effort, consistency across touch points, proactively resolving customer needs, and collecting great insights.

These objectives highlight the importance of a true omnichannel experience, as opposed to a disjointed multichannel experience, and proactive support.

On the other hand, contact center professionals will not be prioritizing some of the traditional success metrics of days past, such as reducing call volume and reducing handle times. Instead agents will be empowered to focus on providing the best possible experience for customers, no matter how much time that interaction takes.

To read the full report, which includes a plethora of additional data, download here.

Download CCW Market Study


Kustomer Queues Make Routing Painless for Contact Centers

Kustomer is happy to announce our brand new, enterprise-grade queueing and routing feature. Queues are essential to managing high-volume contact centers, and for good reason. Queues enable contact centers to monitor their inbound traffic in real-time and optimize customer wait times. Kustomer has adopted this model for the omnichannel world. Now organizations can create queues according to their business priorities across channels, assign queues to teams, and have full visibility into the real-time status of queues and agents.

Unique benefits to using Kustomer Queues are:

  • Queues work in real-time so managers can identify peaks in traffic and reassign agents accordingly.
  • Queues are set up so that every conversation can be in a single queue, so conversations are viewed and handled by a single agent.
  • Wait time, handle time, and wrap time are well-defined so are easier to track and optimize through reports and live dashboards.
  • Agents’ status is connected to the Kustomer Router, thus agents will not get assigned with new conversations when unavailable or during a call.

Queues that work for you

Queues can be defined in any way that makes sense for your business. You can create queues as simple as channel queues (like chat or email) or create more sophisticated queues like “Customers with Lifetime Value of more than $10K”, or for “Conversations with a ‘Return’ tag”. Each queue is assigned to a team, so the Kustomer Router will route new conversations to an available agent with that team. As conversations can only exist in one queue, view collisions between agents are eliminated and productivity increases.

Availability-based Routing and Reassigning

When agents start their shift they switch their status to Available to start getting conversations from the queues assigned to their team. If agents go on a break or finish a shift, they can toggle the status to the relevant unavailable status (e.g., Lunch, Break, or anything else). That will indicate the Kustomer Router to automatically stop sending new conversations to these agents.

The agent status is clearly indicated in the platform, so fellow agents can reassign conversations and managers can monitor and reassign agents, based on their availability. This is especially useful when managing remote teams.

 

Multichannel Routing for the Omnichannel Contact Center

As an omnichannel platform, Kustomer Router includes a multichannel routing capability ensuring that when an agent is on a call, the router automatically stops assigning new conversations to this agent (either voice, email or chat), and resume once wrap-up is done. This capability is available, for example, via the Amazon Connect integration.

Real-time Contact Center Insights

Live dashboards provide managers with real-time information on the status of different aspects of the contact center. This is especially useful when managing remote teams. Such aspects include:

  • Agent status – How many agents are available, what agents are working on (overall and down to the conversation), and what their workload is.
  • Queues – How many conversations are waiting in different queues, and are some queues more busy than others.
  • Customer experience – The average wait time for customers on different queues and channels, and the average handle time once they interact with an agent.

Dashboards are updated as frequently as every 15 seconds, managers can even display them on a big screen in the office.

For more information about Queues and Routing and how to implement them in your own organization, check out this article or request a demo below.

 

Deliver personalized, effortless customer service.

Request Live DemoWatch Video Demos