202: Fireside with Sharad Khandelwal

Charlotte Ward: 0:14

Hello and welcome to episode two hundred and two of the Customer Support Leaders Podcast. I'm Charlotte Ward. Today, welcome Sharad Candleval for a fireside chat. I'd like to welcome to the podcast today, Sharad Candleval. Sharad, it's lovely to have you join me for the first time. And you've brought to me the uh idea of a fireside chat, which is um which is something we'll get into in a second because it's a topic that I am really deeply considering from my own team and my own uh organization right now. So I'm very keen that we talk about what we're about to talk about. But first, perhaps you'd like to introduce yourself.

Sharad Khandelwal: 0:59

Yeah, thanks a lot for having me. I'm Sharat Khandelwal. I'm the founder and CEO of Centisome. So Centism is all about helping brands leverage their customer support conversations to improve the overall customer experience.

Charlotte Ward: 1:15

Okay, interesting, interesting. So leveraging those conversations is a challenge that a lot of um service and support teams face, isn't it? And isn't there just so much in there as well? There at the the depth of insight that we can get from from those conversations is is amazing, isn't it? What's your experience there in the types of things that we could begin to think about when we talk to our customers and record it?

Sharad Khandelwal: 1:42

No, absolutely. I think it's kind of goldmine of data, to be honest. And as a founder, like most tech founders, you you're trying to solve a pain point and you're frustrated with. I think that's one thing which frustrates me that brands are sitting on this gold mine of data, but very few are leveraging them. And and and and and even those that are trying to leverage, they kind of struggle because of the volume of data, because basically of the manual process. And and just talking about the gold mine, uh, it's like it's your customers coming to you on an almost on a real-time basis and telling you this is what they are feeling or this is the problem they are having, or even with some positive feedback if they have, right? So like that's what uh kind of frustrates me. Why are you not leveraging this? Why are you relying on surveys? Why are you emailing people uh a month later or a quarter later asking how was the experience? Like you use use these insights. So it's it's it has got everything about your product, about your brand, about your end-to-end customer journey. There is no other data as insightful as customer support.

Charlotte Ward: 2:47

I I couldn't agree more. And and even in fairly low volume organizations, there is a quantity in there that, as you rightly say, is kind of unrivaled by whatever any customer is willing to tell us in that little square box at the end of a survey, right?

Sharad Khandelwal: 3:05

I know.

Charlotte Ward: 3:06

Um so so how do we like how do we begin to think about if if I was thinking, you know what, I've got this Zendesk back catalogue of conversations, and um I don't know where to start. I don't know how I begin to extract what my customer experience is and how my customers feel about the service we're providing. Um, maybe like where I am now, I don't even have a CSAT program in place yet. That's something we'll be looking to develop over time. But what can I do like right now?

Sharad Khandelwal: 3:43

I think the first step is start reading every support conversation, right? Right. As a manager, as a team leader, as a product person, as a as a CEO, right? You need to be aware of every support ticket, right? So let's let's assume the volumes aren't manageable. So the first step is read and tag them, right? What customers have talked about in those conversations, how's their sentiment, right? Are they talking about product or services or or or anything else? Tag every ticket after reading them. And then you have got all these insights from your support ticket. You can run your daily or your weekly meetings based on these insights, right? That's I think that should be the starting point. But if the volumes are not manageable and if your agents are busy, if you don't want to distract your agents in manually tagging, then comes the question of how can you leverage technology, I'm sure, which we'll talk about it later.

Charlotte Ward: 4:38

Yeah, yeah. As soon as you said read every ticket, I could just I could feel the like the tension rising, maybe the anxiety, because we are fairly low volume compared to, you know, we're not thousands a month, we're not tens of thousands a month. And yet it is hard. I mean, I don't know a support team in on the planet that has like a lot of spare time and agents who have a lot of free time and leaders who have a lot of free time to deal with even relatively manageable volumes. I guess the hope you the hope is that you get to every one of those conversations or at least a reasonable enough sample of them, right? But but it does become unmanageable quite quickly, I would imagine.

Sharad Khandelwal: 5:20

Yeah, it it does. I think that's where technology and specifically, something like AI, or even if you want to go more specific, natural language processing can help you automate this whole process of analysis, tagging, and reporting of your tickets. And I think and I think that's where kind of uh our product comes into picture, where when the volumes are unmanageable and when you don't want to distract your agents, you use technology to automate all that process. And so that you are doing in real time rather than waiting at the end of the day or or week or month, depending on your availability before it's too late, actually.

Charlotte Ward: 5:54

Yeah, because then it it almost becomes subject to the same problems as a survey, doesn't it? If yeah, if if my agent has forgotten exactly what happened with that ticket, unless they read every word, they're not going to really capture it accurately. And and uh, you know, we want to save time, don't we, when we're tagging. And um and I think that's a that's definitely a big challenge. So so when when we think about tagging then, and and I've talked on this podcast before about like tagging, tagging taxonomies and things like that, and approaches to tagging that try and capture different aspects of the custom, either the customer sentiment or the customer journey or customer pain, maybe you're trying to capture kind of product pain, things like that. Um, should we should we try and do all of that in one go with our tagging, or should should we be much more focused if we're doing it manually? And then and then I think if we are focused, this is a really long question now. And then I think if we are focused, does AI allow us to expand on that?

Sharad Khandelwal: 7:01

Yes, absolutely, very good point. I think starting with manually gives you the opportunity to understand what are the limitations, right? What you can do, what you cannot do. Like you just touched on, it's not possible manually to tag every aspect of what customers have talked about. And so that gives you an idea of what are the areas kind of you are not covering as a result of manual tagging, right? And and you also touched on accuracy. I think that's again a can be a big problem because tagging manually can be very subjective. It depends on how well trained the agent is. It just depends on that time of the day, how the agent is feeling and what state of mind the agent is to be able to tag that accurately, right? There's so many dimensions and factors which come into play. But when you are kind of leveraging technology, it can be more consistent, it can be predictable, and it can tag across different dimensions, not just uh from a contact reason perspective, but it can also pick sentiment, it can pick emotions, it can it can say whether it's this is about product or service or agent, and it can go far more granular actually than that. You can you can build up a almost a hierarchy of taxonomy with with the technology with manually doing it, it can be very difficult.

Charlotte Ward: 8:14

Yeah, yeah. I I really like that idea of um uh of being able to draw different views on a ticket because all of the views are there in your in your tagging rather than just relying on the few features that you've decided to pull out in manual tagging. I think that that that's really powerful. Um so it's reducing time, it's increasing the sophistication of of our tagging and our ability to pull insights from it. Um is it is it doing like is it doing anything else for us? What else are the benefits, would you say, of having uh an automated approach to tagging like this?

Sharad Khandelwal: 8:56

I think a big one I would say is enabling cross-functional collaboration. Because the problem you got at the moment is support is just seen as a kind of firefighting department, right? You're you're handling issues, and the tagging is also from that perspective, just trying to understand why customers are contacting so that you can you can kind of uh build a reporting or you can reduce the the tickets actually. But as we touched on earlier, it has got so many insights across the entire customer journey. If you kind of make these insights available across the company, then you have a completely new way of improving CX. You are not just uh relying on NPS surveys or or some social media post to understand what customers are talking about. I think that's one big problem which I where I see with organizations, these insights are just limited to support department. This needs to be made available across. So that's where again technology can help because if it's tagging from a different dimension, it's in real time, and you get a dashboard, you can let everyone in your organization use these insights to drive change and improvement. Why limit this just to support people? So I think that's one big uh I would say advantage, which is often overlooked and is just limited to in terms in terms of trying to understand the time saving and accuracy. Yes, those are the benefits, but the biggest benefit is uh collaboration.

Charlotte Ward: 10:18

Which can only come and which only has really deep value if you're able to apply those different dimensions across the organization, as you said. So even if I was being particularly diligent as a support leader and trying to capture product pain that I could then feed through to my product team, I can still only do that in a relatively limited way manually, can't I?

Sharad Khandelwal: 10:41

I mean, yeah, absolutely. And it just becomes more and more complicated if your organization is quite big, actually. The bigger the organization, the siloed it will be. It's more difficult to kind of pass on these insights to different teams and relevant teams. That's where, again, uh kind of uh technology can help you. And then yeah, sorry, go.

Charlotte Ward: 11:03

Oh no, sorry, I was gonna I was gonna ask, what's your experience of um organizations accessing these insights? Does it increase collaboration between support and these other parts of the business? Or is there a danger that by providing those insights in a very one-directional way that we almost increase siloism? Because we say, here's a dashboard, that's everything you will need as a product team. Um, we'll just keep talking to customers over here and supporting them, and you go and take the take the product knowledge aspects, and and we don't actually need to collaborate ever again. What's your what's your experience there?

Sharad Khandelwal: 11:49

Yeah, I mean, I guess it just depends on the organization and the culture there, right? And what you're trying to do. If as an organization you are really serious about customer centricity, right, uh reducing support tickets, then you would kind of facilitate collaboration. It's it's not just about kind of giving a dashboard to a product team, it's about how you can work together, how you both can be on the same page, because the biggest problem is different teams are using different KPIs, they have a different view of the customer, they have a different view of the issues, right? What support thinks is very different than what product thinks. So, how do you bring everyone on the same page? I think this is how you can you can bring everyone when you are sharing the same taxonomy, same metrics, same insights on a daily basis, then you will be like, let's do something about it. Like, how often are we going to look at the same problem? For we have been looking at the same for the last six months, twelve months, let's do about it.

Charlotte Ward: 12:44

Yeah, yeah. Yeah, you're you, I guess, I guess it does obviously need the culture to develop in that collaborative way. But but then when you are all talking about the same data set and you and you are and you do all have access to the same insights, and um, nothing is a surprise to anybody, then, isn't it? It's not like something has suddenly bubbled up out of support and is being presented to the product team that's something that's been a problem with our customers for six months. You know, we we it really is much more fluid.

Sharad Khandelwal: 13:13

Yeah, it is. And I think another factor is which we often ignore is that you as support leader or product leader, most of like most of the people would know about the problems a company is having, right? But the problem is there is not enough data to validate your hunch, right? It's always about I think or I know we have these issues. What what this can do is basically give you the exact measure or the extent of the problem you're having, and since when you are having the problem, whether it's going up and down. So it's like it gives you context, it gives you the exact measurement impact, and and also basically the risk of not doing something. Otherwise, we are all just relying on our gut and hunch. Oh, we have this problem. Yeah, I heard the support team talking about it, but this happens just in November or just in December, those kind of things.

Charlotte Ward: 14:02

Yeah, yeah. I really like that validating your hunches because I think uh in my experience, one thing that product teams often do is form plans for sort of research and investigation that that is based on those hunches. And I think that can kick off like a like maybe a big research project in product or something based on one person, one person's unvalidated hunch, right? And I think that uh even even if you are using hunches to drive the initial investigation, being able to validate those hunches before you go too far down that road, it is it's such a time saver, and time is money, right? So I think it yeah, yeah, that makes complete sense. Um, so we've talked a lot about the value of this. Um I know there have to be some challenges. What but what what are the what are the what in this domain in terms of like applying AI to customer conversations do organizations find challenging, do you think, particularly?

Sharad Khandelwal: 15:13

Um so I think uh you you mean from a tech perspective or I think so, yeah.

Charlotte Ward: 15:19

I think primarily, I mean, because to me it feels like that I this is a pretty new technology, right? And I think that um this isn't something that I had exposure to five years ago, certainly not 10 years ago. So um how do I like as a support leader, first of all, how do I sell the need for this internally to my business? But also what are the challenges those any organization might experience in realizing the value from this technology?

Sharad Khandelwal: 15:51

Yeah, I'm glad you touched on this point. To be honest, this is the biggest challenge uh for a for a kind of a tech startup like us and also for support leaders because it's a new technology, right? It just takes time to convince people about the value, right? So let's say if I'm I'm I'm kind of uh convincing the value to you, you will have a similar challenge in selling this internally. So I that's that's I think the biggest challenge, and I guess that's the challenge with any early tech product actually. It takes time, and that's where you need to find early adopters of technology who are more keen and and kind of open to such products, and then the word spreads actually. But yeah, that that I would say is one is definitely the biggest challenge. How do you make people aware of the value? Sometimes they they they they know it, they can they can see the value, but you know how so kind of support works in an organization. It can be so tricky to get the budget for a new technology when the budgets are fixed, right? To convince you. Yeah, I've never come across.

Charlotte Ward: 16:57

I always say I've never come across a support team with too much money.

Sharad Khandelwal: 17:00

Exactly, exactly. I mean, and and that's that's something which also again frustrates me a lot. That my why should marketing have all the budget? Why not your support team?

Charlotte Ward: 17:12

So is that part of is that part of the internal sell we should do then? Is that you know that that actually this is going to be valuable to the business because we are gonna be able to surface the same kind of potential in our customers as as we are, as any marketing team or any customer success team might be able to.

Sharad Khandelwal: 17:32

Yeah, absolutely. Uh absolutely. And I I guess it's always about proving value. And I think that's again going back to the earlier point. If if companies start initiative of maybe just doing this manually, right? And and showing the value this data has and basically the potential of it, if if they were to leverage completely, they were to leverage all the channels. You can start some simple by just uh maybe tagging your emails and and showing the value. And imagine what could happen if you start doing with chat, your voice calls, if it's your social, if you bring in all the data together, the value it can add to the organization, to the product teams. So it's uh I think that can be one way of uh value-based selling where you show it actually instead of just talking about it. And once everyone is on the same page, then you can say we have proven the value, but we cannot do it manually because it's it's it's it's not just possible. That's where we kind of need technology. Either we we build it internally, or let's look for the product which can do uh which can do this for us.

Charlotte Ward: 18:34

Yeah, and I guess that's what your your product does, right? I mean, we we wouldn't be here talking about this if it wasn't a topic you were passionate about. So tell me a little bit about Centism then in the in this context, particularly, like the kind of value add that we're talking about.

Sharad Khandelwal: 18:50

Yeah, so so basically centim is about leveraging technology to help you automate the repetitive processes within support. Uh, when I say those processes, I'm I'm mainly talking about two things. First is leveraging insights from that data and then automating workflow. So one simple example is what our product does at its core is it will tag every support conversation in real time, whether it's email or a live chat or voice call or social with different attributes, right? Once you have it, then you can leverage our product or a dashboard to understand what's happening across the company. What are what's what are the main reasons driving contacts? What kind of issues are going up or going down, whether it's product issue, operations issue, right? So that's that's again a big win immediately. You deploy this product. And then the second uh win for you can be when you start leveraging all these tags to automate your workflows, like prioritizing your support tickets, delegating your tickets to the right teams, or triggering your macros, right? Uh there it's it's just like unlimited possibilities once you automate. These kinds of tagging of your tickets. And you don't need to build a new product or buy a new product. Once you have these tags automated, you can leverage your own system, whatever you're using, be it Zen Desk or Fresh Desk or Salesforce. They have these capabilities to prioritize tickets, delegate tickets based on tags. You can, again, there's a lot you can do by just automating the tagging of tickets.

Charlotte Ward: 20:26

That's strangely something I'd never thought about. We tag tickets all the time, the same as ever the same as every other support team out there. And as you said, we have a bunch of automations that do all sorts of things based on those tags. Very, very intentionally targeted to react to those tags. I had just, for some reason, I'd never made the connection between automated tagging and kicking off those workflows. But everything else we've been talked about, I've been thinking about it with the the frame of like bringing the customer voice into other parts of the business and almost almost forgetting I can drive operations off it as well. Of course. Yeah, that makes complete sense. Um okay, so I I think then that I'd I'd like to just kind of think about um rounding out our conversation now to something that I um I know is is is going to be a challenge. It's something that we've touched on before in this conversation, which is which is just about that internal cell about evangelizing about this. When I um when I have a change that I want to make internally, um I I think as you rightly noted before, finding those effectively those early adopters, your your internal champions who are kind of on the journey with you quite early on and get it. I think that's a key part of this, right? And and of any change, actually. I've talked talked recently on the podcast about managing change, and I've talked about these internal champions. But I think there's a this is bigger than managing change. This is this is a big kind of an evangelical piece. I keep using that word, I don't know, I'm stuck on it today, but but this big selling piece to the whole organization because I'm doing more than just making an operational change with my team or or adding something that I think will benefit customers purely from a support focus. This could be potentially very far reaching in the business. So um you talked about showing some early value in some small ways from some of this thing. Is there a big kind of statement of value you'd like to finish with that we can all take away and say, you know what? I finally get this thing. This is actually the umbrella, the umbrella statement for AI and for automated uh like extracting customer sentiment automatically from my conversations. What's the big umbrella statement for you at the end? Well, why should we do this? Put you on the spot a little. I'm sorry.

Sharad Khandelwal: 23:34

Yeah, that's okay. I think I would say if you uh move to automation, uh, what you get is basically a 360 degree view of your customer, their customer journey. So basically it becomes a single source of truth, which to be honest, you you I mean, you have a long career. I'm I know you may you must have heard this hundred times. Every company wants a single source of truth, right? But uh there is a very simple way to get to it, is just automate this tagging process. Once you have automated, like the same technology can tag your social, your support, your surveys, your voice calls, and then every conversation, insight from every conversation is under one platform, right? And that can be used across the company. So you are actually as as if you as a support team, you take the initiative, you are adding a lot of value to different teams because they are all struggling. They don't know what's happening in other departments. You are giving them something which would be like, oh wow, I didn't know my product had uh kind of these issues. So it's it's it's like adding value, bringing everything together. That's that's like a single source of truth for everyone in the company, not just leadership, but for everyone.

Charlotte Ward: 24:44

I really like that idea of like the single source of truth, the 360 360 customer view really in action. Um, and from a support team where it should come from, right? Thank you so much, Shannon. Thank you. Thank you for your time. Thank you for joining me. Do come back and have another chat.

Sharad Khandelwal: 25:01

Absolutely. Thanks a lot for having me.

Charlotte Ward: 25:07

That's it for today. Go to customersupportleaders.com forward slash two zero two for the show notes, and I'll see you next time.