Prakhar Vats is a founding member of LimeChat, which leverages conversations in WhatsApp and Instagram to double your ecommerce sales. With their innovative AI technology, LimeChat is able to be seamlessly implemented into chat platforms and has created a productive way to advertise brands to customers and increase sales.
On this episode, we discuss conversational marketing, the importance of linguistics in this new marketing strategy, why messaging channels will be the primary buying mediums in the next five years, and much more.
What is Limechat
Prakhar Vats: Right. We're a conversational AI company. Very simply put, we help e commerce businesses sell, market, support, to their customers on chat platforms. That platform can be any platform, you know, outside the US. WhatsApp is pretty big.
So WhatsApp is one of our major, you know, WhatsApp is one platform for which we are optimized, but in the US, we're also doing it for Instagram, for your website live chat, for your TikTok.
Alex Bond: And when you say your website, live chat, that little kind of like Zendesk style chat that a lot of people see on the website.
So that's interesting to me because, you know, again, in the US, I know some people use it, not everyone, my cousin that lives in Ireland uses WhatsApp, but why did you choose that as the primary platform for your service?
Prakhar Vats: It has a lot to do with where we started. So I think we started with India, you know, we're a company that was started in India. WhatsApp here is the platform of choice.
So, you know, I think the other day I was reading, reading some statistics and it said, you know WhatsApp has fairly amazing penetration into the Indian market and the average WhatsApp user, right? The average WhatsApp user in all of Asia, I would say spending on average, almost 50 minutes per day on that app.
That was, I think, indication enough for us to kind of take this platform seriously. All the e commerce brands that we work with. They also requested that, hey, you know, we receive a lot of queries on WhatsApp. So if you are building an engagement tool, if we are building something that will help us use a tool, use a communications platform to sell more, to market better, to handle support queries better, it should be WhatsApp. That's what we did.
Alex Bond: That's very interesting to me because there are so many of these other types of apps and channels and things like that. While WhatsApp is probably the number one service internationally, I was just curious.
So when you're initiating, can you kind of run me through the process at least of what a client can expect when they come to your company? If say a client wants to use your service as a brand, what specifically you do for them and what they can expect from your services?
Prakhar Vats: Right. And just say two words, it's chat commerce. So if you run an e commerce brand, you already have a lot of people who either visit your website or if you run ads, when you ask people to message you, people already come in and they have a lot of queries.
Hey, you know, what's the material like, when will I get it right? I happen to have an engagement, you know, three days from now, would you be able to deliver it before that date? We just take all of that. We automate that. You send a query, much like chatgpt. You send a query, you receive an adequate response.
We're able to track, you know, usually brands get these queries around. Hey, where's my order? Please cancel it. Things like those, you know, is it good for my skin. This is a product that's certainly for dry skin. Will it suit me? Things like those we automate on chat. It's basically like when a customer converses with a brand, but they're giving away a lot of information on chat, right?
So we build these personas for these brands that, Hey, this is the customer, or this is, this is one of your customers who would benefit from a campaign like, you know, I like this, maybe like a black Friday or a cyber Monday campaign. Maybe another persona where somebody openly mentions that, Hey, I'm not really into discounts.
They only have really, really in depth questions about the product. We tell them, Hey, you're better off sending this one time promotion to a client or to a customer who is not of this type. So we help you with basically, you know, again, to use those two words, chat commerce. So think of it like this.
What's your favorite D2C brand? What's your favorite e commerce brand these days?
So let's say you want to buy from Nike on WhatsApp. So you don't have to open the website. You don't have to go to an Amazon. You have Nike in your WhatsApp. You just message, Hey, I'm looking for something like this size eight.
Maybe, you know, maybe give it to me in black and you get a list of options. You select and you pay and complete that entire transaction within chat itself. So you don't have to leave chat to, to make a purchase right now. This is the first part where, you know, our AI helps Nike or a similar brand automate these queries.
The second part is. Now, Nike knows Alex now, right? Because Alex has just messaged Nike. So all communication that you'll receive from Nike from this point onwards will be optimized. We'll have this memory in mind that, okay, we had this conversation and Alex likes this.
The next time, let's say there's a release of this, there's a rollout of a product that's similar to the one that you asked for. You'll know about it. And it's all happening on on the chat platform, so it's basically that.
Alex Bond: And one of the big advantages to it being totally and completely involved on what's app is there's less opportunities to drop the sale, essentially, I mean, especially when you're on like a laptop or a website or.
You start on Facebook, then you go to a different website, then you go to a third website to actually pay for products sometimes. And it, it just becomes to the point where I don't know what I'm doing here. And I don't know if it's worth it, but when you're all on WhatsApp, it feels so much more legitimate. Was that part of the goal, the objective?
Prakhar Vats: I mean, definitely. So I think I would say, I basically say that what we've found out is. When somebody messages you think about it, they have a pain point. And it's these high intent users that message you on, on a platform, like let's say WhatsApp or even a live chat, right?
I asked people, hey, when was the last time you went to a website and you clicked on that Zen desk? I mean, probably, I mean, I've never done it. So it's basically, you know, these are people with high intents, right? These people want to buy from you. These people have a query that, that they basically see it fit that, Hey, I'll take an extra 30 seconds out, I'll type this message and send to this brand.
I mean, e commerce brand would love somebody like this, wouldn't they? Right? This is somebody who wants my products or is interested in my products, right? Why do I not reach them out? And if you can have a channel that does not perish, right? And why I use the word perish is let's say I go to Nike's website right now. You know, maybe tomorrow.
I'm not somebody who goes through my previously open tabs. I think I have more than 1000 open right now. I'm not going to go through each one of them. And then, okay, well, I was in this frame of mind.
Maybe I was looking at size age, black sneakers, maybe something for my cousin. But if it's WhatsApp, it's all there, right? So if you receive a message, you'll just scroll up and say, Oh, yeah, I was looking for this, right? And okay, I've Nike sent me this. So that's the advantage of, you know, having that conversation on chat.
Alex Bond: It seems like a major advantage. How do you determine who your target audience is for a brand? What are some of the things that you look at if a brand wanted to engage a person vice versa, the other way around?
Prakhar Vats: So if a brand wants to engage a person very simply based on when people buy or when people message you based on, let's say, you know, you're somebody who only buys during, let's say Black Friday, Cyber Monday, or maybe when there are discounts going on, right?
Let's say you're a discount digger. Some people only buy when, you know, a fresh stock arrives. So it's just personalization, right? And, and we built on top of these, these logics that brands already have. So if you ask a brand that, Hey, I mean, do you know who your discount diggers are? They probably all know already.
What we do to that is we add conversational data. So we're basically a company that analyzes chart information, chart history, and then we're able to get to that. Then we're able to. And I figured out, okay, who will be, let's say most suited on a channel like WhatsApp. You don't want to send out too many messages in a given day, right?
Because people, people don't like receiving these messages. So let's say you only have a hundred thousand messages for a given day. LimeChat tells you, okay, these are the 100, 000 people you should send to for maximum conversion. So that's also something with them.
WhatsApp Marketing: How brands Initiate conversations and drive sales
Prakhar Vats: The very first thing is obviously not sending it to somebody who has not consented to it, right? That's obviously the basics. What we recommend brands do is that, Hey, do it transactionally in the sense that somebody has done something on your website.
Let's say it has to be personalized per se, right? So let's say if I go to your website, I check out all the options and then, you know, I click notify me on this one product only message them, you know, when that product is back in stock, right? It should add value to the customer.
So all the flows, all the automations that we've built somehow, you know, guide a brand in this direction that, Hey, just because you have, let's say 10, 000 odd people who have bought from you before, you know, you should not pick up that list and send a broadcast to all of them.
One is consent and two is obviously adding value to the client. So these are the basic, I think it's not too different from email marketing. If you, if you think about it. Right. It's just that email has less views, less click through rates, et cetera, et cetera, right? It's not some, it's not a place where somebody spends 50 minutes per day, right?
Very simply put. It's very similar to email marketing, just that it's much more active. It's much more personal. So you have to be even more aware of, you know, when you're sending that message even more aware of, Hey, do you have the consent to send this message? Things like those.
Alex Bond: And it's more live. It's more dynamic. People are usually messaging every 30 seconds compared to an email, which is like 30 minutes or something like that. So you kind of got a Bob and weave in real time a little better.
Can you automate specific responses depending on how a customer is engaging with the chat. I mean, how does that process actually look?
Prakhar Vats: So it's very, I would say, obviously, simple would not be the right term to use for this, but.
Alex Bond: Simple to you, maybe not to me.
Prakhar Vats: Yeah, but very simply, it's basically we've built an intent detection model, right? So think of it like, you say things like, hi, hello, how are you? Right. That's all intent.
Great. So, you know, we basically, I understand that you're somebody who's greeting, right? Let's say you go to a certain product and then you ask questions about how to use this. What's the material like? So we understand, okay, this is the intent is to know more about this product.
So a lot of it is just realizing or figuring out what you're asking about and going to your website and then, okay, checking and then responding back on chat itself. So yeah, it's pretty simple. The most tricky part I would say is when, let's say you want to automate your order cancellation.
Somebody's placed an order and then they text you, Hey, I want to cancel it. Right. And you know, for you to automate that. You need to have certain logics built into the system. Do the check whether it's been shipped or not. Do the check. Okay. If it's, let's say a high, high value order, let's say it's greater than 300, 400. Would you not want to cancel it instantly?
Let's say if it's a small amount order. So all these checks and balances, you know, we've built into the system. So brand can just come in, set up all of this. And then, you know, start using automation, so to speak.
Strategies to prevent customer frustration
Alex Bond: And I'll tell you from a personal experience that I think one of the things that frustrates me most when I'm specifically chatting with an automation or, you know, with a robot essentially, is that it takes me so long to talk to a person to solve my problem that it's trying to be solved in a way that I'm not getting the solution that I want, or I'm not getting the questions that I'm asking answered.
And it takes a little longer than I'd like to finally get to speak to a an agent of some sort. How do you prevent from customers or potential customers from getting frustrated with that process and thus not wanting to use that brand at all, especially since you're kind of more. I mean, you're obviously in the retention business as much as the conversion business, but how do you kind of prevent that from happening?
Prakhar Vats: I think, and this was something that we dove deep on almost three years ago, that why does this happen with automation experiences? You know like you mentioned, it takes you so long, so long for you to reach to a human. The more fundamental question we asked was, you know, why do you actually need to reach to a human, right?
Why do you actually need to talk to a human? That's something that we figured out and why most automation tools fail to give that experience is because their horizontal system. So by horizontal systems, I mean, they'll work for banking. They will, they'll work for e commerce. They'll work for, let's say, maybe, you know, a restaurant, right?
They're built to, in a way, enable, let's say the business to create their own bot flows to create their own automation, you know, I would say journeys. So, let's say as a restaurant owner my entire life, I've been frying steaks or maybe, you know, all my life I've been, you know, tossing up books, you know, I don't know anything about making the perfect automation flow, right?
So I have to rely on what this automation tool is telling me. So naturally I'll not be able to make an experience that's impressive. So what we realized really early on is that if you have to truly kill that experience, I mean kill in a good way. And if you have to truly make a kick ass customer automation journey, we'll have to be a vertical company in that we only work in, you know, one domain and within that domain, everything we have covered.
So we picked e-commerce because obviously e-commerce growing as ever. You know, we will continue growing. So we kind of built a system that was, you know, built, optimized, and, and planned for the e-commerce use case. So that's how we kind of, you know, got covered ourselves on that trend.
Retention and first-time conversions in WhatsApp
Prakhar Vats: Oh, my, I think I can go all day long on this part. I'll give you broad numbers. So we are experimenting with like I said, right? If you run ads where you invite people to chat with you So in India, we're calling it like India, Southeast Asia, Brazil, right? We're calling it CTWA. Click To WhatsApp Ads. Where you show somebody an ad and you don't say to them, Hey, go to my website and check these things out again.
And then Google pixel or Facebook pixel will track you. And then when you send you, when we'll send you an email and then we'll show you these ads again on Instagram, just tell them, Hey, if you're interested, just drop us a DM. We'll sort it out. Right. And then you're able to figure out, okay, what product are the truly interested in.
It's like an interactive journey on chat. And obviously there will be drop offs, but in this case, it's personalized and it's in their DMs, you know, you can follow up and you can, you know, how we like to call it is up the ante, right? Obviously this is what brands do. So you go, you show interest in a certain product, you know, obviously you drop off, they'll reach out to you, you know, in their minds, you know, more and more creatively to give you more discounts, to kind of make it more easy for you to purchase.
Because this is a straight interaction. There's no, let's say an ad involved, you know, there are no, I would say cases involved where you have to go to Facebook again, to see that again, to see that ad again, or maybe you have to go to Instagram again, or TikTok again, to see that ad again, where you initially, you know went to the website through it's on your WhatsApp, it's on chat itself.
You can just simply go and then revisit that conversation. If you receive an update, you know, exactly what you discussed the last time. So the ROAS numbers that we are seeing is 2.7x a minimum, right? So. If you're spending 100 and you're getting a certain amount of revenue, you'll be getting 2.7 that if you happen to do that same on, on, let's say click to a DM ad, maybe I wouldn't call it WhatsApp, but click to a DM and that this can be Instagram DM as well.
So that's the kind of conversion numbers we are seeing. The secret sauce here is very simple. A compelling first time user. journey, right? And to a really strong follow up personalized follow up system. So these are basically the two things that go into the secret sauce to these higher conversions in terms of retention.
Obviously, we have to put like a definitive number to this, but it's generally, and again, when I speak of generally, I mean, from the very time the person has bought, you know, they're in your retention journey on WhatsApp. Right. For these customers compared to, let's say the ones which are not in a WhatsApp journey, right?
Retention is 30% higher and CLTV is almost 44 to 47. I mean, depending on, you know, the kind of brand you also run, but it's in general 40% higher than let's say normal conversions that happen through email, WhatsApp, email, or the websites. That's basically you know, WhatsApp is capable of, and not just WhatsApp, any conversational medium, even Instagram DM for that matter.
Alex Bond: We'll definitely talk about the Instagram DM stuff. I think that's very interesting. That's something you guys have been cooking for a little bit now. And generally when you're comparing the efficacy of your work, it's this a click to DM versus email and SMS. Is that kind of what you're trying to compete with or measure up against?
Prakhar Vats: For click to DM, it's like somebody views your ad maybe for the first or second time, and then they're coming to talk to you over there. The comparison is straightforward in a straightforward way. It's with the website conversions.
So you run the same ad, let's say same creatives, let's say maybe a slight middle of the funnel type of an audience, which somehow knows your brand has interacted with your ads before, you know, when we ran AB tests, conversions were basically 2. 7x of what you saw.
In a primitive, I would, I mean, I call it primitive. Like it was a long ago, but let's say as what people have been used to it, you know, you invite people to the website. So the ROAS that we, that we're able to generate is 2.7x for the retention journeys, which I mentioned where, you know. Somebody is carefully pushed into an automation floor or a retention journey, as we call it, right? A user journey is again, is comparable to SMS and email.
Determining when to stop engaging with customers
Alex Bond: So how do you determine when to stop engaging with a specific customer or subset of customers to avoid being marked as spam?
Prakhar Vats: Right. That's an awesome question. So first of all, very simply, I mean, with every, so we maintain it, like this is. I would say really good practice that if you are messaging somebody, or if you're conversing with somebody as a brand on a, on a personal medium or a personal platform, like WhatsApp, or let's say Instagram DM, you should always give them the option to tap out of that flow.
That's something we send with every message. So in case you don't want to be part of these alerts going forward, just. Click on this button. Let's stop. That's where 40% people drop off after the first message itself. So it's basically telling them, hey, hey, I don't have a problem with your brand great product, but just keep me out of this constant pinging on WhatsApp, right?
And that constant, I mean again, this entire depends on how the brand wants to set it up, but we don't recommend doing it any more than twice a month. Right. Because again, it's a personal medium and I'd be okay with that. So that's something we recommend to brands. This is the proactive way of dealing with this.
The reactive way is obviously to figure out, okay, how does my quality rating look? So on a, platform like WhatsApp. Meta really carefully manages every every interaction that that, you know, if you are a business account on WhatsApp, every interaction that a customer or a user is having with you, they carefully manage it.
So let's say a lot of people report you as spam or mock you as spam, your quality rating would. Start to decline. So the second or the reactive way is to, you know, just if you're, if you're familiar with online gaming, there's this thing called spray and pray, right? Where, you know, you're in a panic situation.
You just press on the shoot button and you, and you hope for the best. You do that. And then you obviously based on the feedback from the controller, you kind of control, you know, how much. More bullets you're willing to fire. So that's the reactive way of doing things.
Alex Bond: I think that's a good analogy. And I like to hear that, that you usually recommend no more than twice a month, because I am, I I'm thinking in my head, oh, these brands are contacting like once a week, every other day. And I would just be like, get out of my hair, honestly, but twice a month, I probably get contacted from like SMS or email stuff.
At least some of them I asked for every day, but that's because I'm really interested in like those opportunities that they're providing and that's more like newsletter stuff, but like an actual brand contacting me, I think two a month is actually really interesting. It would definitely pique my interest more often.
Prakhar Vats: Just on that point, really quickly, I think why two a month. We'd also work here is because let's say, you know, you go to a certain website, you check something on your drop off, right now you receive two messages. You know, one of these messages looks like, Hey, Alex, this happened, you know, just click this link. Buy, buy, buy.
You know, again, we're not open to having a conversation with you, right? This is the first message that you receive. The second message is, hey, Alex, we know that you were doing this. Now, there are just two buttons, explore further, unlock a special discount just for you, right? We're gamifying that experience.
We're telling you, hey, I mean, if you click on this button, who knows what happens? Even if you send like two messages per month of this kind of the second kind, people will click on either of those two buttons. And then again, they get added to a conversational flow. So in that moment, they can, you know, inquire more, they can ask about different products, things like those. So that's the whole idea here.
Utilizing WhatsApp commerce
Alex Bond: How does LimeChat help Shopify stores in terms of utilizing WhatsApp commerce?
Prakhar Vats: I'd say in three ways, right? One is we help you make a catalog on WhatsApp. Now what does that mean? You are having a conversation with a brand. You're able to access all of their products on the right. That's as simple as that.
So while you are, you know, whilst conversing, you can check all their products, right? That's, that's the first thing. Second thing is obviously building these automation flows. So, you know, we help these brands set up support automation journeys, you know, pre-purchase journeys where, you know, like I said, where somebody comes in and they interact, ask about a product or maybe.
Ask things like okay, you know, is it suitable for my skin? You know, what's the material of this cloth like things like those? We help them with that and thirdly it's you know, basically helping them set up these user journeys on chat.
So post purchase, pre purchase, in the sense that you've bought a product. Now, what should be the next five messages over the next two months that should be getting delivered to Alex? So we help them design all of these things as well.
Instagram vs. WhatsApp
Alex Bond: So you've mentioned it a few times already. And I want to dive into kind of the Instagram side of things. So your newest major endeavor involves working with meta to integrate line chat into the instagram platform What are some of the differences or obstacles in working with instagram versus whatsapp?
Prakhar Vats: That's an awesome question. So very simply like think about this. How many websites have you seen where let's say it's an e commerce website And they ask for your number versus how many websites do you see where they ask for your Instagram handle? Like probably it's like a hundred percent versus zero percent.
Nobody's asking for your Instagram handle, right? So that's the first challenge where to create, let's say an automated journey for let's say a post purchase journey, right? On Instagram, it's impossible, virtually impossible because nobody gives you that data. I mean, and brands have no interest in receiving that data.
Let's say, even if I start asking for my customer's Instagram handle on my website, right? What if they change it to something else? You know, people usually do that on Instagram potentially lost this customer forever. So that's the first thing where due to this reason and this reason alone, you know, we're not able to kind of build a platform that allows you to send these alerts to customers by basically.
Having a checkbox on your website that okay, I consent to this, right? Because obviously you can't do that So that's the first challenge where retention becomes really tricky on the retention side but you know working with meta and and obviously meta has rolled out a really awesome feature I'm, not sure if you're aware already.
It's called you know communities where If you happen to follow Mark Zuckerberg on, on, on Instagram, he sends out, he sends out these weekly or daily updates on these communities. So whoever becomes a part of that community, whoever becomes a part of that broadcast list receives all of these updates on DMS, like I said, to counter the fact that, okay, nobody's going to give you their Instagram handles on their website.
We help brands become these broadcasters and ask people to join their communities. The Instagram algorithm till now, you know, it was really messed up where, you know, brands reach 1 million, 2 million followers, but at the end of the day, you put out a post on the, about a hundred K people are watching it only about 100 K people get it in their feeds, your loyal customers.
Or there's no sense of community. It's like a Pandora's box, right? You know, you never know what's going to come out of it, right? In the sense in the sense that, you know, you're not too confident about it. We help them, you know form these communities over Instagram. So let's say as a brand, let's say Nike, you know, to continue with your, to continue with their original example, let's say, you know, you want to subscribe to these alerts from Nike.
Let's say every time there's a new product that's launched, you can just simply go to their profile and click on that button. Subscribe to this. That's how we're solving the retention or the, or the one time the promotional side of things, whereas the marketing that's involved on the automation bit where people come in and then they ask questions and then you have to automate WhatsApp or Instagram or even live chat, you know, people just come in and have two kinds of questions, either questions about products.
Or questions like, Hey, where's my order? You know, can you track it? Can you cancel it? Things like those. So that's already, you know our, our engine is already capable enough and, and is, and is deeply integrated with your Shopify store. So people ask these questions, we give them this data instantly. And it's like interacting with the human almost.
So yeah, on the automation bit, it's not too different. I would say only the kind of queries are somewhat different on the retention bit. It's definitely more tricky, but Meta has rolled out certain updates that help with Instagram. The biggest use case that we came across was common automation. If you had a brand with, let's say a hundred K followers, you put out a post very easily, you'll be getting a thousand, maybe 800, maybe 700 comments.
You know, on every post that you send most of these, you know, are maybe, Hey, can you let me know how much this is for? Price, please. Right. Et cetera, et cetera. People, people just send you these comments, right? What if I told you there was a way where you can pick up, let's say I am a brand and, you know, let's say Alex comments on, on Nike sports.
You make a comment and then instantly like he sends you a DM saying, Hey, Alex, you asked about the price of this particular product. This is what it is, right? You can use this, you can click on this button to buy. So comment automation becomes even more important.
So the biggest use case that we saw, and this was something that was requested by every brand was, Hey, I mean, I get a lot of comments that I can't, I can't automate them because, you know, people say a host of different things and any automation flow that I can build, it'll include at least.
One response and then multiple options. You know, I can't put options in Instagram DM because let's say even if I were to put options, you know, what if they respond again? So that's that's not an interactive medium. That's just a comedy. Leave a comment. And then it's a comment, right? So that was the problem.
But, you know, we built that tech that allows you to take, you know, let's say if, Nike has live chat on board anybody who comments on Nike's posts, like price or material or things like those, Nike would be able to just initiate a DM with them automatically.
So now you've kind of picked this person who has just commented and you have taken them into an interactive channels, which is Instagram DM. That's something that ties up with the first thing that I told you about automation flows, but yeah, that's how it works.
Alex Bond: I really appreciate the well thought out answer. And frankly, I think the commenting automation, the response to comments, automation service you provide is super innovative.
It's definitely not anything that I've heard of and I'm extremely interested how would determine what's worth automating or following up on, you know, for example, if Nike posts a picture of a new pair of air maxes or something like that. And I say these look cool.
You're probably going to have like a million comments that all say Awesome, looks great, wonderful, and even negative comments, you could automate this to persuade people to get more information about it. So I'm curious kind of what you do to determine what's worth following up on.
Prakhar Vats: That's an awesome question. I get that a lot. So at the outset, you take a look at any Instagram post, let's say by Nike again, 30 to 40% is just fire emojis. Yeah, I mean, awesome draft, fire emojis, fire emojis, fire emojis.
What we did was when we started working with these brands, we kind of figured out, okay, what are the kind of comments that are the most redundant that, that brands get fairly frequently. These require nothing. You know, you don't need to react to them. You don't need to respond to them. You know have to do anything. Fire emojis and things like those.
We kind of added into this list where, you know, the bot was not supposed to do anything. There was no conversation or DM that had to be initiated. Right. Then came keywords in and around price or costs and things like those. For those, it was somewhat understood that, okay, this user wants to understand more about this product and the price of it.
So the DM that gets initiated says something like, Hey, Alex, you wanted to know more about this product. Here are the details. So this is the price. This is, you know, the material, things like those. And you can, again, Click on certain buttons and know more about that product.
Negative comments, a very big part of it. There was something that was also very heavily requested. People say, you know, my past orders are delivered. You had to maybe, you know received bad service over here and here, right? Things like those people comment that stuff.
You just initiate a DM with them and say, hey, we realized you have not had a great experience. Please let us know what to do. And just two buttons call our toll free number, or maybe, you know, again, in the case of. Nike, maybe, you know it'll be so huge that they wouldn't want to do anything.
They just want to get them to just respond in chat, give the right feedback and then leave. But for a small brand, this means a lot.
So for a smaller, for a smaller brand, you know, the severity of the negative comment, we can then, you know, initiate a DM and then say, hey, what is it that you want to do? Do you want to explain your issue? Do you want to get in touch with a human and things like those?
Alex Bond: You're doing PR or damage control in real time too. I mean, I think that's a huge aspect to it that not a lot of brands are able to do. You can't just like hear what everyone's talking about and be able to say, hey, here's maybe a misconception about what you're thinking about or what you're saying about, and I think there's a lot of value in that.
Prakhar Vats: Absolutely. Yeah. I think with some of these Shopify brands that we're working with, they have North of a million followers on Instagram, right? Like I said, a post you put, I mean, even if you put a blank picture, you're going to get roughly 1000 to 2000 comments, right? Even if it's a blank picture, just take a picture of the sky and just put it.
I mean, 1000 people would will will comment, right? So, and a lot of that would be again, fire emojis or maybe, you know, something like that. Some of that would be here. I got this amazingly, like, pathetically, you know disappointing service, things like those. And then some people would be, you know, some people would actually have like fairly long comments that you would not be able to understand.
Like I said, with the engine that we designed, where we just put it into three or four buckets, where, you know, one is no reaction, two is, you know, more about the product or more about the subject in the picture. Third is obviously, you know, it's a support or it's a PR thing, right?
Where we have to do damage control. It's the tone is that. And fourth is again, if it's a long one, it's not the first type, but it's a fire emoji, but it's none of the other types also, you know, nobody's talking about the product. Nobody's talking about, you know, nobody's talking about bad experiences.
Let's say you post a picture and somebody says, hey, long time ago, I was in the US, I was doing this and that makes no sense. He just initiated a DM and saying, hey, what do you want to explain that better? And then obviously, once they respond, that chat flow initiates from there. So, yeah, these four buckets, I would say, are something we designed.
Keeping up-to-date with AI technology
Alex Bond: How do you, with being so heavily dependent on AI and AI functionality, how do you continue to implement that sort of tech that is evolving at such a rapid rate? You know, I mean, I could imagine it requires a lot of updates, but it is moving at such an exponentially quick rate that I could imagine if you take your eye off the ball for a second. You're in the dust a little bit, you know.
Prakhar Vats: Absolutely. I think what's beautiful here is that end of the day, the functions or the actual steps that, you know, maybe like a logical conversation is to be followed by a physical action in most cases, right? So let's say you have a discussion or let's say you have a conversation with a prospective customer about.
You know product they'll either buy or they won't buy so With all these advancements these changes, you know, the latter part or maybe the actual physical the action part is not changing so if somebody wants to track an order the process would remain the same if somebody wants to cancel modify place You know reject an order the process would not change what will change.
However is you know, the words that we use to get You to do that, right? So we just have to kind of, you know, continually advance on those fronts where every day our model gets more and more trained because, you know, we're handling over, you know, 10 million conversations on a given day.
So we know what people are saying, you know, what conversations are about products, what conversations are about support.
So we're able to train our systems better on that front. I think, you know, that's where advancements are required, but you know, the core actions and this thank God we chose to be only an e commerce focused, you know, automation tool that's definitely in our control because the four or five or maybe even the 10, 20 things that get classified as actions remain the same, right? Have remained the same for the last two to three years.
