Pivoting is not the end of the disruption process but the beginning of the next leg of your journey. This is exactly the essence of the journey of our next guest on Founder Thesis.
In this episode, Akshay Datt speaks with Raghu Ravinutala, Co-founder and CEO, Yellow.ai, a hardcore techie who has worked with Texas Instruments and Broadcom before charting his entrepreneurial path.
In 2014, he met his now co-founder, Jaya Kishore Reddy, who was then, working on a product that aimed to bridge the gap between customers and enterprises, as a weekend project. They soon quit their corporate jobs and took up this part-time project to a full-time daily mission, founding Yellow Messenger in 2015. Today as Yellow.ai, it is one of the World’s Leading Conversational CX Automation platforms, which is trusted by more than 700 global enterprises.
Tune in to this episode to hear Raghu speak about how Yellow.ai aims to democratize artificial intelligence by offering chat and voice bots to automate customer experience for large enterprises.
What you must not miss!
- The transition from B2C to B2B.
- Developing a multilingual platform.
- Experience of working as a part of the Microsoft Accelerator program.
- Fundraising journey.
- Plans for global expansion.
Raghu, what is home for you? Where are you from?
I’m from a place called Kurnool in Andhra Pradesh. Not sure if many of you heard,
Is it a village or a town?
It’s a town. In fact, it’s the first capital of Andhra Pradesh before Hyderabad became the capital. So it’s a pretty good town.
Tell me about your journey through the corporate world so to say.
If I look back, they were the best years of my life, right? I think somehow, I just feel that these corporate years that I’ve spent actually turned out to be the most productive, most fun and most exciting years of my life. And at Texas Instruments, when I started, I had a very good boss. I think that determines, I mean, how your work life is actually around. And I think that was a place where the concentration of talent was amazing, and a pretty flat structure, even in 2000. And there was a lot of intensity around innovation and intensity around solving technical challenges. And yeah, that was a fantastic experience that’s quite memorable. And I think that motivated a lot to learn more about technology, how people manage technology, how technology companies are run, which is quite different from what you see how traditional industries function in general. And from Texas Instruments, I joined a startup at that particular point of time called Magma Design Automation, which was into software to help chip design. That was my first tryst with startups, stock options, equity in a significant manner. Okay, because I think we were at Texas Instruments evaluating software and had chosen Magma to be the provider. And so Magma stock was rising in front of my eyes. And so while it’s a great opportunity, I was one of the first five-six members of Magma design automation in India.
Magma was an Indian company?
No, no, it’s an American company. The founder was again, an NIT Surathkal guy, but had a pretty good India presence in general. And that’s a company where I went through the entire process of working with enterprises, spinning enterprises deals.
From a techno-commercial role at Magma, where you were also involved in winning deals into a purely technical role at Broadcom, like, did that make you feel restless? And was that the impetus to eventually be an entrepreneur? What was the impetus?
So the impetus was never dissatisfaction with whatever I was doing. I think it was an exciting journey. So there is nothing that was lacking, let’s say in any of these roles, all of them were challenging. All of them were very nice teams. So there’s no dissatisfaction. I mean, to be frank, I could see the startup sector coming around. Entrepreneurship, by itself, had an attraction and even at Magma, and even at Broadcom, it was a company that was acquired. I joined the team that was acquired by Broadcom. So that entire journey of building something from scratch, and actually becoming very rich was very fascinating. And I felt that that’s an experience that was worth giving a shot. Right. So I think that was the impetus rather than I mean, anything to do with the satisfaction of any of the roles.
What was the first step? I’m assuming initially while working only you would have started exploring ideas. Tell me about that journey from being an employee to being an employer.
So I was always looking for ideas in general. So I had 5-6 ideas over a period of time that included a sports social network. And they didn’t include it. They were speeding any commerce deals, etc. And out of the ideas, the filtering process that I kind of approached at that point of time was out of these ideas, which one is something that I can sell to any business in the world. And every says a consumer could be a potential user of the platform. So that way, the market is pretty big. And there is a good opportunity for being a successful company or product.
You wanted to have a B2B product like you were clear on that?
So when we started, it was not a B2B product. It was a B2C product. In fact, I wanted it to start as a B2C product. The reason was, we started the company while I was still working at Broadcom. And it’s very tough to do a B2B company while you’re working somewhere. And B2C was the fastest way of reaching out to the market. But I think the concept that we thought about was that there’s so widespread adoption of WhatsApp in India. And if there was a way for consumers to interact with businesses also in a similar channel, that’s something that every business would potentially benefit from and every consumer would benefit from. And there was nothing like that at that particular part of time. So we actually launched a B2C app that listed businesses and enabled consumers to start messaging businesses. So that was the first avatar of the company.
When did that decision to go full time happen? Like you initially started a B2C Just Dial style kind of platform. So how did that evolve then?
So we almost worked part-time for like one year, and we were already seeing that we were being written about, investments were approaching us. There were already about 25,000 plus transactions happening for days, people were downloading, reviewing, good reviews and all that stuff.
You had an app?
Yeah, we had an app, we had like 100k downloads, even when we were working part-time, three of us.
What was driving these downloads? Like what was the use case that people were using it for?
So it started off with people trying to book restaurants, with people at that time, even Swiggy and all were new. People even ordered food quite a bit through this. We also had recharged, etc. So that serves multiple things, the initial downloads.
How would it work? Like you could search for something and then you would have a chat button?
Yeah. If you search for, let’s say, a nearby restaurant, and you click that button, and then somebody says, I want to order something, we have like two-three member call centres that would respond on behalf of the business and actually make that particular order or transaction work and provide customer support for that.
Where is WhatsApp in this?
There’s no WhatsApp. It was our app which was like WhatsApp for business. Okay. So somebody downloads, let’s say, Yellow Messenger. And they choose any business, it could be even Flipkart, or Amazon or it could be a nearby shop and starts chatting about what they want from that, just like they would chat on WhatsApp. So there was no WhatsApp at that point of time. It’s our own app. But it kind of functioned like a WhatsApp for business.
How did it become what it is today? Like, what was the next version of the product?
As soon as we came out, we said we needed more visibility, more support. So we started applying for accelerators. We went to Startup Chile, we got some 40k dollars from there. And of course, I think, some visibility. Then we went to the Microsoft Accelerator. So what happened by that time was that what happened to be our idea in general, the world was waking up to this. And Facebook started to talk about it, Microsoft started to talk about it. And Facebook Messenger opened up for businesses, suddenly, there was a lot of need, or hype about messaging with businesses suddenly, like conversational commerce, that term came in. So when we started, there was no term. So by that time, we had automated these conversations, and we had a pretty good tech build-out by the time the hype started to take off, right? And then Microsoft said, our customers are asking for it. We were part of the Microsoft accelerator, which was introduced to the Microsoft sales team. And we were like a four-member team at that point in time. And they said You know what our customers are talking about this. They’re asking if Microsoft has a solution. We don’t have a solution. At that point in time, we can introduce you to those customers. They started introducing us to Tata Power, Asian Paints, IOCL and all of them were pretty impressed with what we were showing them. I think we were the only company that had even a working prototype at that particular point in time. So they started giving us some deals.
But this was all for Yellow Messenger itself, not for integration?
No. So we changed the model, right? So we showed Yellow Messenger. But what we said is that the same interaction can happen on your website, the same interaction can happen on your mobile app rather than our mobile app. Because we found out at that particular point of time that growing a B2C company in this space is going to be challenging, that too when we are bootstrapped. So we don’t have the entire ecosystem of businesses and all that, we’re a small team. So let’s go and try to sell in a SaaS model for businesses. And at that time, we’re still focused on B2C. Oh, once we get all the businesses, we can still do a B2C path. But let’s go and work on B2B websites because that’s going to give us revenue immediately, we’re going to integrate with our systems, we won’t even seek users and customers without us spending on customer acquisition. And we said yes to those customers. And that’s how the journey started. And once we started the journey, the customers actually were seeing good RoI. I mean, once it launched on their websites, the number of phone calls to their customer support was going down, their consumers were happy, and their CSAT scores were increasing. Then we got Bajaj Finance as a customer, and they were expecting about just 50,000 users to kind of adopt the platform. But within three months, it has gone to almost a million users per month. It’s just like the product-market fit was so strong, that the first 5-10 customers that we saw of them, expanded their consumer base on this platform. So we clearly landed on a clear product-market fit that if you have automated messaging-based customer support on your websites and mobile apps, you’re going to see your customer support cost reduce and your customer experience improve, it was a proven thing. Now we went out and built a platform around it on how to create these conversational journeys, how to build analytics, how to do all the enterprise features that are required to scale to multiple enterprises. And now we had reference customers, had a good platform, we had a partnership with Microsoft, then we started scaling to more and more customers and adding from there on it kind of took off, right? So we bought a million ARR.
What does an NLP API give you?
So at that time, the NLP API was giving something called intent and entity. So intent is like, you can define what the customer wants and train it based on a few utterances. An entity was like, you can kind of think of intent as a verb and entity as a noun. That’s how they started. So you can understand what the customer wants you to do. And also the noun on which you want to do that action. That is what NLP API was starting to give at that particular point in time.
Okay, and what are your pricing is like, was it based on the number of transactions? Like the number of conversations?
Yes, number of conversations. That’s how we were priced right. Even right now we are pressing based on that.
When did you hit the 1 million ARR?
We started selling this model in 2017. At the start of 2016 also was sold. But in 2016, we were selling more like one time payments, right in the sense that you know what I mean, build deploy this cost extra, because we never knew whether this will kind of work or not. But once it started working, we started on the SaaS model in 2017. So by 2015, it started with zero ARR and in 2018 October or so got to 1 million ARR numbers. So it took about 18 to 24 months from zero to 1 million ARR.
And Indian customers?
Yeah, all Indian customers, the spread of little bit of customers from the UK, but the majority 90% were Indian customers,
And how did you build the sales organisation? Because enterprise business means you need a strong sales team.
Till we hit 1 million ARR, I was the only sales guy in the company. There was only one more that I hired from somewhere else. And yeah, so the majority of those deals were done by me.
How did you crack high-value deals? You must have learned how to do sales? Or was it something which because of your Magma experience you were good at?
I think that played quite a big part in what I would say. I think though in Magma I didn’t sell I was in close contact with the sellers who are selling really high-value deals. And frankly speaking, right, I think our business model was quite right at that point in time, I think glad that we chose the model that we chose. So we were really not closing high-value deals to land in the sense that I think most of the customers would have ended up paying pretty less as a starting fee because it’s based on conversations. But where I think the low lands actually expanded to become high-value deals is when the number of conversations expanded. So from let’s say a 25 to the almost million conversations automatically, the deal value became 200k ARR and 300k ARR. So I think sales just got us into the door. But I think the real value and deal expansion happened when the users of the product expanded. So the way that was automation works as somebody calls from the phone, either inbound or outbound says I need to let’s say, I need to get a personal loan and software understand what the person is talking about provides options and starts the conversation either loads the deal or provides the response back to the end customer.
This sounds very, very intensive to build out like you would need something which will understand like, which will be multilingual.
Absolutely, I think that’s why we have a 130-140 party member software team building this out, right? So we have folks from Google, we have folks from Myntra and Flipkart, they have PhDs. So it’s a pretty large engineering team in the company that builds out these products.
And it has multilingual support.
Yeah, absolutely, I think we are the world’s number one in multilingual interactions because we have customers in Malaysia, Indonesia, Hong Kong, Japan, Brazil, Mexico, and of course in India. Multilingualism is an absolute requirement, and no other company in the world is processing as many multilingual conversations as we are doing. So yeah, absolutely.
So your first fundraise was that small friends and family round of one crore. So tell me about the funding journey subsequent to that?
Yeah. So in 2018, I think when we were closing on one million ARR, in fact, we were hugely profitable. We had 40% profit margins at that particular time. So funding was a pure choice in the sense that they were really contemplating whether to raise or not because I think as for the company’s survival, we don’t need to. And we went to this program called Extreme Entrepreneurs from Lightspeed Venture Partners, I think they had a pretty good program. And they had these global startup leaders talking about their experience and stuff. And then we went to that program. And of course, as a part of that, we shared our story, our product, the company. And then Lightspeed expressed interest in investing in the company. And that point, have we just talked about in a way that if we take funding versus not take it, which one would we repent more? And we felt that, yeah, I think maybe at this point of time, it’s kind of quite a good scale. Lightspeed is a great partner from what we heard. They have a great brand name and great companies, and we really liked Dave Caray. He was the investing partner. So we went ahead and accepted that CDC fund of about $4 million at that particular point of time to help us kind of build out our sales team and expand into new markets. So that was the first round, I think, the Salesforce accelerated when we built out the team, etc, and in a year we grew by almost 4x, from where we are. And then we happened to meet the Lightspeed US team that visited India and Lightspeed India introduced us to them. And they came back saying they’re the kind of traction that we are seeing. And the kind of global customers that we were having, it is something that is pretty exciting. And they’ve seen a great market for us to expand in the US and beyond. We had interest from their team, and that helped us close on our second funding round for $20 million.
The second round was essentially to go global?
Absolutely. What happened is, just before we wanted to go global, we had COVID. The second round was during COVID. Actually, while we kind of closed the term sheet before COVID. By the time I closed the transaction COVID was pretty big, and that was lockdowns and stuff. So we delayed our hiring in the US. But we built out a team in India itself that kind of targeted global markets. And COVID was a pretty significant tailwind for the company. We pivoted our messaging. WhatsApp became a key channel for enterprises to engage during that time. So we grew another 4x during the one year after. So we moved from over 110 people to about 550 folks over the last 18 months and 4x growth in ARR. This year, we felt that having the capital markets are at their peak and a great environment to kind of raise further money and expand even faster. So we raised our Series C I think just a month back or so with some of the best partners in the world. We had Safar Ventures which has had about 20+ IPOs in the last five years, WestBridge Capital which has some of the best companies in India, US Corridor and also invests in public companies. And we got the company that has the best understanding of customers across the world with Salesforce. It was pretty great to get Salesforce investing as well. So that’s been our, like funding journey in general.
Are you at liberty to tell me the valuation?
We haven’t shared the valuation. I would say it’s pretty close to half a billion dollars.
Are you now hiring on ground people also in these global markets? Are you looking at India-based?
No. So we have a pretty reasonable US team. So right now I’m in the US, I don’t know. So we have like an 18-member team in the US. And we are in fact, having a US geography sales kick off in New Jersey in the second week of September. And our plan is to get to a 7200 member team in the North American market. We are hiring for local field sales in every single market. So we have our team in Brazil, we have our team in London, we have our team in Mexico, Singapore, so pretty much out there. We have a sale set in Japan and Thailand. So absolutely. We are investing a lot in sales and marketing in local markets.
What do you think helped you outpace your peers, like say, you know, as you said, Haptic, which has been around for a much longer time and has the backing of Reliance. But you know, what’s the secret sauce?
The secret sauce is one, I think, being aggressive in the market. I mean, you need to build out your marketing sales teams, so that you’re in front of a lot of customers. Second, I think we were very agile. I think both started as B2C companies, I think we moved to the B2B model pretty fast. And we work with some awesome customers to start with. And we had a tremendous amount of focus towards making them successful, be it our work with Bajaj, be it our work with Asian paints, be it our work with Procter and Gamble, etc. So we learned a lot from those deployments. And the product improvements are essentially driven by the scale that we were seeing with our initial customers. This product became way, way superior to any other product in the market out there. So any discerning customer, I mean, it is very tough for them not to choose us, the only reason they potentially don’t choose us is potentially on the pricing. But the technology because of working with phenomenal early customers kind of put the product in a much superior position. So it’s a combination of I would say strong product and aggressive distribution.
What do you see as your fastest-growing markets globally?
Yeah, the fastest growing is definitely the US market. So India continues to be a very, very strong market. But the next huge growth will be seen in North America.
And that is driven by more and more businesses, opting for automation and customer solution,
Yes. Automating customer support and the platform has expanded too so that companies are adopting yellow AI even for HR automation, IT automation, and the product has been built to deliver multiple use cases. So built on infrastructure where people can configure conversation use cases or self-serve platforms, and the ease at which they can deploy for multiple use cases. I think that has grown. And we have some significant partnerships with Microsoft and of course with Salesforce, that kind of help us get to the market faster. So that is driving a lot of growth. And of course, we built our sales team as well right now. So that’s like a determinant of how much you can sell as well.
And between the channels of communication like say WhatsApp or voice or cloud, telephony and say Facebook Messenger and say your native chat application inside the customers’ app. Which of these channels is likely to be the fastest-growing?
Yeah, right now the fastest-growing is voice. So that’s because of the lower base as well, I would say, but right now it is the fastest. The second fastest-growing right now is WhatsApp. So it is our own websites and mobile app-based SDKs. But overall, as a volume, the web and mobile app have the largest volume because they were the first channels that we kind of deployed into the market. And this seems comfortable growth but it’s like new channels are opening up fast and growing.
How do you navigate around not promoting spam? Because you know, like, say, people are getting WhatsApp marketing messages, and then it becomes seen as SMS. You know the way today, I hardly open my SMS app because 95% of messages are promotional. So how do you navigate that?
I mean, we don’t do that right. What we do is depend on WhatsApp and what the companies want to do. So WhatsApp has pretty stringent guidelines around what can be sent through outbound notifications. And there needs to be consent from the users before sending those.
How’s that consent audited? Is it based on the company saying I have consent? Or is there anything else?
No. The advantage of two-way platforms is you get feedback from customers. If you see any message that you got from WhatsApp, it says, Is it or intended you wanted? Do you want to put it in spam? More people put it on spam, there’s feedback, right? And the account is automatically blocked, deactivated, right? So it’s not like, as some theoretical audit. SMS doesn’t have that other option, right? So is it intended or not. And even for outbound calls, there are clear TRAI guidelines, etc. And we see ourselves as a platform as an enabler. We don’t, because we are not sending the messages, we are giving a platform to send conversational messages. And it’s up to how the companies that are using us determine what is the right way to use the platform. And they bear the consequences of, either way, they can use it in a great way that improves customer experience, where you’re not waiting for customers to ask you, but you’re proactively informing them information, versus using them for spam. So we see ourselves as enablers than the marketers, we are not the marketers, we are the providers of a platform that provides a phenomenally great way of running outbound campaigns, which always have been one way without any feedback to actually being two way. Enterprises realise that that’s a superpower that they can use to amplify. But yeah, it can be used for spam as well. But I think it’s up to the enterprise, we don’t want to be the controllers of that,
You know how email as a communication channel today is super accessible for even one person freelancer. If I want an email automation platform, it is extremely low cost for me to set that up purely as let’s say a podcaster doing user engagement. So when do you think messaging will reach that kind of accessibility? Where you know, it is like super available, super low cost, just self-service? You know, because as of now, from what I understand, WhatsApp is expensive, especially if you compare it to email as a channel of communication. How do you think that market will evolve?
So I think even right now, WhatsApp can be done self-serve. So we have launched cloud.yellow.ai. So you could go get a WhatsApp account and start messaging customers.
How much would that cost?
Right now for interaction, it is about 29 paisa. Yeah, I think you can get started with that. I know that it might be a little more expensive than email. But this is what I tell people, right is that nobody sends an email, nobody sends a WhatsApp message for the sake of sending a message, there’s always an outcome that you want to achieve. The outcome could be read rates, the outcome could be conversions. The outcome could be lead generation. So you can use a low-cost channel, get five leads versus you paying something and get 20 leads, it’s what delivers you the impact for your effort and for the money you spend is the channel that you want to kind of use because nobody sends SMS for the sake of sending SMS. So if your read rate is very low, there’s no point and especially in India and emerging markets, how many people even access email? Yeah, for you, it makes sense for podcasts. I think people on the other end of the spectrum who use technology a lot, have emails. In India, everyone has WhatsApp, but very few have email, right? So there’s no option other than to use that. And second, if the interactions are initiated by the end customers and not proactive, then it is free to use. If you have a great customer base that engages you on WhatsApp, it’s absolutely free to use. So that’s how I look at it.
So what is the role split between the three co-founders like who does what?
I do primarily on revenue sales and marketing. I take care of that piece. Kishore takes care of the technology. The entire engineering is run by Kishore. Rashid takes care of the product and Rashid is the only person in the company who can write code and who can sell software. So he connects the entire loop.
Are you now based in the US to drive sales there?
Drive sales and build the team here, etc. So essentially trying to build the US market for the company.
What are the challenges of managing a globally distributed organisation workforce?
Yeah, it’s like lack of proper sleep, because different times that’s the number one challenge, I would say. I think that’s the number one challenge I would say is, I think there are demands from different geographies. And thankfully, I think we have some great leaders in the Asia Pacific who are pretty much running the show. That helps me sleep better. But I think how we manage our time energy, I think, is the key challenge of running a global organisation.
Your family has moved with you?
Not yet, potentially next year. So with all this COVID, It’s very tough to move around. So I came in, and I think potentially they’ll move in next year.
What do you see as the end game? Is it an IPO or acquisition? What do you see as the long end game?
The long term end game for us is definitely the IPO. That’s the north star and we are working towards that but anything can happen in general.
Working towards that as in a 1-2 year timeframe?
It can’t be 1-2 years. For an IPO, you’ve to at least be a 200-250 million dollars company. I think it will take us about another 4 years to get there. So a realistic timeline would be 4-5 years.
So that was Raghu telling Akshat Datt building about Yellow.AI. If you would like to know more about this, then do log on to their website www.yellow.ai.