Product Head, CaratLane
Ajith Singh heads the Product team at Caratlane – A Tanishq Partnership. With over 12 years of experience across building consumer products, scalable business platforms and solution selling of IT services, he has been part of Caratlane since it was a startup to becoming India’s largest omnichannel jewellery retail brand. Prior to his role at Caratlane, Ajith worked with companies like Microland, CSC and IBM in providing B2B solutions to large clients.
- [01:49] How do you build delightful customer support
- [05:49] Has the buying behaviour of consumers changed
- [07:59] Is there a need to reimagine customer support
- [09:41] What impact did the pandemic have on customer support operations
- [15:25] What are the best practices in customer support
Joining us today is Ajith Singh who heads the product team at CaratLane.
Ajith has been instrumental and has journeyed with CaratLane and is now here to talk to us about how customer support in this space has been evolving.
Welcome to the podcast Ajith. Why don’t you start telling our users a little bit about your role and your journey with the company?
Thank you Siddharth. Thanks for the introduction. And, as you told, I head the product team at CaratLane. As you know CaratLane is India’s largest omni-channel retail brand.
Right. We started in 2008 mainly as an online brand to start selling jewellery online. We were one of the first in this segment.
Later we adopted the Omni channel strategy, right. So we now have 120 plus stores across India. We do sell across international geographies as well. We have other channels like try-at-home, live video calls, chats, calls and other channels. Apart from… uh… Jewellery being a luxury good, jewellery purchases is often emotional.
The customer wants the touch and feel of the product before buying. Buying jewellery, if you see, is often related to the ceremony or a big emotion, and CaratLane’s aim is to make this customer experience a memorable one.
And coming to my role, as part of the product team at CaratLane, our job is to understand customer problems and the business needs and build… build system solutions to bridge the gap.
We do a lot of data crunching, analysis, and talk to a lot of customers to understand them.
Understanding customers, their behaviour, the problem is a large part of what we do at the product team at CaratLane.
[01:49] Siddharth Sharma
Got it. Ajith, let me tell you something. I’ve been a customer of CaratLane for a very long time and I think you folks have truly redefined the way jewellery buying has been done in India.
The way it is handled and I… I can attest to the fact that you folks look at your customers at a very granular level, because everything is highly customized.
And it is… it’s a very pleasurable experience, right? At Verloop also, we stand for the same thing about… you know, how we build delightful support. I think you folks are doing it very well from a product perspective.
Thanks. Uh…, Good to hear that, right? We are always pleased to see customers of CaratLane talking back to us. And a lot of our vendors also are customers to CaratLane. Happy to know.
Right. So Ajith, getting back into the conversational AI and customer support right, I mean folks… because you folks do a fantastic job.
I mean, in terms of CaratLane, what have been the major customer challenges? What have you seen in terms of customer support?
Yeah, very good question Siddharth. Considering the situation where we are today,
It’s most relevant today. The support teams at CaratLane are traditionally based out of a central location sitting all-together. One thing we did even before the pandemic was diversify the team across geography. So we have teams sitting out of multiple officers now. But the major challenges we faced during the Pandemic was supporting from phone, right, where the teams were not used to see each other.
A typical conversation happened in a flow that I walked to the next… nearby team, asked what… ‘hey what’s happening’ or this is what it is, right?
OK, I’ll go ahead and answer the customer or I’ll patch on the call to you, you take care. So that was the most difficult part of our…part of the challenge we faced.
But the teams were… like quick enough to adapt to the new one…uh…newer way of working. and initially we face problems like calls dropping and chats lagging.
All these were coming up. Then we kind of… solved the infrastructure problem by helping people… for logging a complaint. The next major problem was providing real time updates to the customer.
As I earlier told, being a very small team, everybody used to talk to each other and tell each other what’s happening. Now that’s not possible. They have to reach out to the other one through system.
It brought us a very peculiar problem and an interesting problem to solve. We have to digitize and standardize the way of asking questions to each other. We started creating an internal follow up ticketing system. Now we exactly know what kind of ticket has to be answered by the other team and how much time they take.
Started measuring what’s our SLA and took advantage of that scenario.
In essence, that’s the toughest thing of not meeting people, sitting in silos at home was the biggest challenge.
I think that… that’s a very unique problem to solve, right? I mean, when people are spread out right, I mean, that human bond touch is there and that is just basically distributed. Now how do you maintain it on a digital platform?
And I’m sure over the past few years, you would have seen the industry also evolve. What are your insights into that?
See, I already told, CaratLane was an early adopter of technology and one of the first to sell jewellery online. but… later we understood that the… touch and feel is… major part of what we are doing.
We want to go where customers are. So we opened up stores and we went to customers’ homes, where… our try-at-home service and we adopted the live calling service.
All this started happening. So, recently, if you see, everybody wants to come, digitise their service and everybody wants to be present online, right?
That’s catching up big today. Even in small CaratLane, being a digital-first company, the hardest problem was to make customers come digital, right?
So, a lot of our customers directly walked into our stores, saying, oh, we didn’t hear about you online, but we just saw you, right? So we want customers to discover us first and then… then visit us.
That problem was heavily solved during the pandemic because everybody started going online. Sales happen either through online or different channels, but most of our customers are now very comfortable with online. That’s a major shift which has happened in our industry particularly, in a very niche luxury segment industry.
The shift has… That shift has happened in the last one year.
[05:49] Siddharth Sharma
In… in… in terms of, you know, the pandemic and… bought on a very different shift in the buying behavior, right?
And you’ve just talked about this. What is your opinion on how much automation led by AI… you know, plays a part in pre and post sales phases?
Before talking about how AI can play a role, let me explain about what’s a typical pre sale and a post sale conversation in CaratLane.
A typical pre-sale conversation is all about helping the customer make a choice, telling more about the product selection and which is… telling mo… more about a product which is not listed on the website for them to read and understand from. And other questions on policies- different policies, how deliveries happen and solving payment related problems. That’s a typical pre-sale conversation.
And coming to post-sales, the major questions revolve around the order statuses and how can I do returns and exchanges, or I service my product, what happened to my refunds, right?
Can I get faster deliveries, right? So all this comes under the post-sales bucket. Either pre-sales or post-sales, whatever queries come to us, can be segregated into three cohorts. Where information exists in the system and it exists in digital format, right? One. And information exists in the system and it’s not digital yet.
And information doesn’t exist in the system. I would like to probably cohort 1,2 and 3. Cohort 1…1 is an excellent candidate where AI can play a major role. The information exists in the system as well as it’s digital, and AI can actually very accurately give information as close to a human being.
The next place where AI can play a significant role is user qualification. You see, in a CaratLane kind of a company, the most constrained resources are the sales force and the support staff.
And it’s very important to qualify the user before they start speaking to the support agent. So, AI can play a major role in understanding the customer
Asking basic profiling questions, classify them and then take them to a human who is actually able to solve that problem in a much better way.
That’s where I see AI can play a major role… where information already exists in the system as well as local qualification. At least for CaratLane, these two are major areas where AI can do.
[07:59] Siddharth Sharma
You… you made a very interesting point about the three cohorts, right? Yeah, you know where the information exists, and all the context is there, where the information slightly does not exist is not there, and there is no information exists, right?
I’ll just like to pick on that thread and, you know, you also told me that the support teams are basically the most stressed… one of the most stressed teams in… in… in any organization, right? And… especially when you are looking at a… you know, post-pandemic world.
Do you feel there is a need to reimagine the way customer support has evolved over the years?
Definitely yes. Customer support has evolved over the years. Even within CaratLane if I see, when we started, we started taking questions in form of email for our portal, where customers type in their questions and wait for an answer.
Later on, when the entire queuing method started, everybody started publishing a toll-free number and customers started directly calling. Post then, we needed user qualification to calls to happen, right, and we had the IVR setup
where… you will select 1,2 and 3 and go into the queues directly. Uh… later on, the live chats happened. Every business started putting a chat icon in their website or the app and people started chatting with the business directly on the site. So… but if you see, every business wants to go to where customers are.
If customers are there on social media, I’ll go to social media. I’ll give a real example. Where… if you post a question on Twitter or Facebook, you will get a faster and better response than going into the chat and the email platform.
That’s the reality today, right? Where everybody is living. So, that’s where we are. When businesses started entering it to channels like Facebook or WhatsApp or even Facebook Messenger, that’s the trend going on, right now.
Rather than asking customers to come to the businesses, businesses reach out to the customer where they are. That’s the trend we are able to see in the last couple of years.
[09:41] Siddharth Sharma
I mean, we’re picking on the thread about… you know, where you told about how buyers now are on social media and how brands need to reach out to the customers and the users where they are on their channels.
Do you think Pandemic was a turning point that has forced organizations to innovate when it comes to engaging with their customers, especially in support?
Exactly yes. Actually we are seeing that trend be… getting evolved over the last couple of years, but the speed of adoption was not this much.
We are traditionally a digital first company. So we are first to adapt to any technological platforms, but a lot of traditional organizations who have only an offline presence, we can start seeing that trend happening at a fast pace… in the last one year, right?
Everybody… uh… forced to see. Even in our example, where the entire tele shops were shut, everybody started selling the gold coins or the gold certificates online, and that… that’s the first step. After one year, one and a half years from the start of the pandemic, everybody has a full fledged online presence.
So pandemic was a major face in… induction in the entire adoption process, right, I would like to say that.
Even for us, the way we do things has tremendously changed over the years. We were just speaking to customers and chatting with customers. We didn’t know what was happening.
Recently we have started developing our own AI and ML process to understand what the customer sentiment is like when they speak to our agents. First they are chatting on the phone like, right and need to rate these calls and chats to get more insights out of it.
Internally also, we have adopted even a bit more rigorous AI/ML logic on top of the traditional chatting or the calling part.
Slightly getting deeper into customer support and… you will have a lot of insights on this.
In terms of customer support, right, in just pure customer support, do customers still feel the need to talk to a live, (a) human agent or a… or do you feel that, OK, automating… a good process of it is sufficient? Or is it OK that, you know, customers are adapting to it as well?
Well, very much customers want to speak to a human, at least in the industry where I come from, right?
So, ours is a very niche segment where there is a need for a lot of personal touch. And, it is not like we have not tried AI in the past, we have tried… uh…
We have tried automating certain queries in the past. After seeing the negative customer experience, we have chosen to revert it back. I’ll just give a typical order status example.
If you see what’s there in an order status and… If you see, we have a certain set of information which is already published to the customer in forms of an… notification on an email, sms or he would have my account. Even if customers chat with us or call us, that’s kind of the same information we have. But if customers are reaching out to us
There is something more than the system or the bot identified, right? That part is something that bot was not able to recall for us.
So we just thought that we will redirect the chats to be humans rather than bot. Particularly in our segment
So , as we speak, even after the pandemic, even in 2021, a call centre is very much needed.
And now I just… I mean, we talked about… you know, you talked about the certain limitations when it came to and you had to fall back to a live agent.
In your opinion, what are the present limitations when it comes to… you know, AI driven conversational AI that includes chatbots and voice?
Whatever I have seen, in the… in the industry where I come from, understanding customer intent is something AI and chatbots or even voice bots are still evolving.
Often, we see customers cannot clearly communicate what they want.
If a question is clear and the answer is there, It’s very easy for the chatbot to be involved, but it needs a lot of probing to even understand the actual customers involved and the actual customers problem.
So these are stuff where the human agents still score over the chatbots. And sometimes, customers are very furious and angry over an issue, whatever may be.
And the best thing to do is avoid a chat bot and directly redirect to the human. We have seen that happening multiple times.
If a chatbot replies… or even a human replies that ‘we understand your concern’, that statement is often taken with hostility, saying no, you don’t understand my problem. We have seen customers all react in that way.
One more thing is… uh… AI or machine learning needs a lot of data to get trained and make the module better. In places where the data is less, the accuracy suffers.
So in an industry where our volumes are less and the ticket size is more, AI takes a much longer time to make hits there, and… speaking about CaratLane, we make the majority of selling through assisted channels, where we feed the customers either in person or through a chat or a call.
Even post discovering the product online, customers need some kind of reassurance, at least in the segment of Luxury and jewellery.
And sensing emotional connections is something chatbots are still evolving where humans are very good at. And… CaratLane is a very high experience category. Where the questions and answers are very standardized, chatbots can play a great role and this company feels the support function as a cost metrics.
CaratLane does not see that as a cost metric, but rather as a conversion metric and as an experience metric. That’s where every agent gets rated on CSAT and the experience they provide as well as the conversion they make.
That’s where chatbots still support and humans still win. I didn’t say chatbots and AI have not made a significant progress. These have really taken off.
Within the next few years, I would personally want to see a revolution in the AI… entire AI/chat bot industry. Indian based chat
Bots are still available, but still figuring out, configuring, training as they making the model learn, it’s still a pain… to… just… really know about that, right? And… so, I’m sure these things will get easier and better in the upcoming months.
[15:25] Siddharth Sharma
Got it. And I think you answered my next question as well, when I was going to talk about… you know, how the changes in the way customers interact with the brand can be crucial. And you talked about… you know, assisted selling.
I think we’ve… you’ve answered that question for me. And… I’ll jump on my next question which I had. What do you think are the best practices…
when it comes to providing customers with delightful support experiences, and what used to work before and what is working now?
Sure. Very… very good question actually, because… uh… as I already touched upon, uncovering customers’ real problem is still the most important link in our entire sales function.
It was very important to know that customers often think of a solution in mind and ask for it. Unless you know what is the intrinsic problem behind it, you will not be able to propose a better solution. Maybe there are better solutions available in the system.
But, that’s something our human agents are very specialized with, and we have seen, a lot of examples of going beyond their..uh… regular way of support to make customers experience a delight. I’m sure you’ve heard this example from the Zappos delivering happiness book that…
how they were able to refund the customer even after the death of the customer, and… even when they are not able to find a card has expired… they find some other way to make that happen, right? So, they bend all the rules and… to make the customers experience delight.
And, CaratLane is nothing short of that. There are enough examples of our agent taking the product, flying down to customers place
reaching the customers in person, going to customer’s engagement, finding alternate product, knowing the customers better, going in person to the engagement, giving… making the engagement happen. There are enough examples in CaratLane scenarios as well.
That’s one thing we are… very peculiar about CaratLane. We go all out there… whatever is possible, ok, beyond the rules… defined rules, to make customer experience a delight. Everybody of us, are very sure of.
Wow, that is so insightful and that’s so brilliant, right? I mean, especially when we talked about Zappos, right? They… they are the pinnacle when it comes to, you know, delivering the best customer experience.
So, lot of this examples are talked about internally… within our leadership community as well, and, often you see the top leadership members responding to customers directly. Even… I can see at least one or two emails responded by our CEO himself, right to… directly to customers, even today, right? Even with the scale and volume what we operate today.
That is so brilliant. Alright. I’m now going to jump into… you know, picking your brains on to… asking you about a little bit on the future, right? How do you think customer support will evolve in the next two years? And what technologies will it evolve with?
Good question, but I feel two years is still a small period, OK, but the trend towards AI driven solutions and chat bot will continue taking off in the next two years.
Uh… the more and more… regular, mundane work will be offloaded to the AI systems and… more sophisticated the questions… of answering… completed questions will still happen with the human being.
I would say a sweet… whenever we are able to reach a seamless transfer and harmony between the AI and human is where we reach an optimal level of support.
I’ll say an example, the AI in probing the customers about the questions and understanding their need and figuring out what could be the answer and… say, a person responding is a better fit,
redirecting the question to the person, they giving the personal side, solving the customer problem and, again, giving back to the AI to get the feedback post the support, right, that’s kind of an… harmonized living between an AI and a Human.
I envisage it would happen in the next two years. Everybody will make the world a better place to live.
And… do you see the convergence of voice and chat in AI solutions for… especially for customer support? You see that… you know, your conversational AI for customer support will now start including text and voice in the chatbots?
We have not seen that trend yet because… when you speak about the voice chat, calls are still the preferred mode, at least in the CaratLane scenario than a voice chat-based solution. More than voice chat, we are seeing the trend moving towards a short form of videos or story format right, Instagram… reels kind of a formats.
That’s picking up at a great speed. We feel that video chats… video chats will come a big way at least in categories where… touch and feel… is the problem. Video chat will come a… major way and solve the gaps that are in the system. And… at least, from where I come from, voice chat does not pick up that much.
I am… particularly wowed by how… uh… CaratLane… uh… you know, looks at customer support, and especially… you know, the last part of it about how even the CEO today writes emails to customers, and like, customer experience is at the core of what CaratLane does. We are really glad to have you. Thank you once more Ajith.
Thank you. Nice interacting with you Siddharth. Thank you.