Today’s Scatter Brain is brought to you by AngelList!

AngelList Stack is for startups that want faster fundraising, cleaner cap tables, and high interest banking, all in one place. Scaling companies such as Abound, Harness Wealth and Syndicate migrated from other vendors in less than a week with zero legal fees to gain an unfair advantage in managing their back-office. Learn more by signing up here.

Whatsapp is aggressively enabling businesses to launch native experiences within its messaging interface. You can order an Uber from Whatsapp in selected cities in India. You can also browse, buy, and pay for groceries from JioMart, a joint venture between India’s largest retailer, Reliance Retail, and Jio Platforms, a subsidiary of Reliance Retail’s parent company (one of India’s largest business conglomerates, Reliance Industries).

I talked to Haptik’s cofounder and CEO, Aakrit, who has been building conversational UI-based products for a decade. I have a personal connection with that market. Back in 2016, I used to intern at a venture studio called Betaworks in New York City. We ran an accelerator over the summer for chatbot startups. That was the peak hype period for the emerging category. That also happened to be my first exposure to the world of startups. Apple had launched the iMessage app store. Facebook got serious with the Messenger platform. WeChat was already a big success story at the time, and all its western counterparts were trying to emulate chat-based experiences.

In this conversation, Aakrit and I talk about :

His journey tackling conversational experiences
Reflections on the B2C play and the pivot to B2B
How failure can be a privilege
Unspoken sources of conviction for founders
Conversational UI as a design problem and the broken discovery paradigm
Balancing sticking to a strategy and not avoiding opportunism
Culture of Jio
Bangalore versus Bombay
Underrated opportunities in India

Sar: I watched a video of your pitch for Haptik at the Mobile Sparks conference in 2013. You asked the audience how many of you hate customers calling customers. And pretty much everyone raised their hands. The first version of Haptik was a conversational experience in a consumer app aggregating customer support channels for a bunch of companies you partnered with. The current form of Haptik is a B2B company. Do you look back on that and say, “thank God we don't have a consumer play anymore?”.

Aakrit: I was out in Silicon Valley recently. I met with an advisor with whom I have had a relationship for 15 years. One of the things he mentioned as part of how we should think about building things was that we should do something where, even if you fail, you won't regret the journey. I had so much fun during those four years of my life that I focused on the consumer app part of Haptik’s journey. That’s also when I met John at betaworks, and others focused on chatbots. Everybody was trying to figure out what this was going to look like. Would it be a mobile app or sit on top of a messenger? I don't regret it. I think if there's anybody who tried the hardest in the world, it was probably Haptik. And only by the end was I completely convinced that the world would not look like that. This will never be a solution that the world would ever need. We pivoted to becoming a B2B company. I'm happy I gave it my all, but I also think I should have committed to B2B sooner.

Sar: How did you build conviction around chatbots in the first place? The platforms were still quite nascent in their developer ecosystems. There wasn’t as much chatter about conversational interfaces when you started in 2013.

Aakrit: I worked at a Silicon Valley startup called Flurry in 2010. We used to make tools for app developers and were one of the earliest analytics providers. At one point, more than 100k apps were customers of Flurry, and a large category was messaging/communication. Apps like WeChat and WhatsApp were part of this customer list. So, as an admin to the tool, I saw some crazy engagement on messaging in general, more than any other app category, including social and gaming. This made me realize there is something here - messaging can become the interface of the future. So I started dabbling with ideas in the space and landed on the large narrative of Business Messaging: if it’s so easy for people to talk to each other, why can’t there be a similar app for people to talk to businesses and get things done?” That was the epiphany.

Sar: When I think about conversational products, I think there are two things. One is the context, and the second is the experience. You haven't given up on the conversational experience to this day. What you gave up on was the context of a consumer app. Now, the context happens to be the messaging interface. Is that how you look at it?

Aakrit: Can messaging or conversational experiences become a better way to get things done? That was always the first line we wrote when we started. That was always the mission. In that sense, you're right. Today, we're solving the same mission but very differently. But I would've loved a world where it was done in a consumer app. When you try and do it through businesses, you are one degree separated from where the problem lives. Fundamentally, there's limited control. You are trying to solve end users’ problems, but your customers are the companies.

Sar: There was a lot of tourism in this market. Almost everyone I met in the ecosystem in 2016 has moved on. VCs that made massive bets in this market have moved on; some have taken down blog posts about how conversational interfaces are the future. You stayed in the game long enough to capture the value once the market took off. Did you come close to giving up mentally on this idea back in 2013-2016? What role did your ability to raise capital play in staying in the game?

Aakrit: I am okay with the thought of trying and failing and getting up again, but I'm super paranoid of failure. I feel like we glorify failure too much in the startup world. I didn't have that luxury of affording to fail back then, which made me paranoid. I had to figure out how to turn Haptik into something. I think that's what kept us going. I often think about this: whenever I start my next company, will I have that same thought process because now it's different?

Sar: There's a smaller chip on your shoulder ten years in.

Aakrit: Yeah, I was all in. I didn't have any other option. I started it with my cofounder when I was 27. I got married when I was 30. We had a kid when I was 32.

Regarding your point about fundraising, I feel you have to have enough conviction in your actions. A conviction can come from three places. It can come from your love for a problem statement, and you will not sleep without solving it. It can come from not having any other option because you will suffer in terms of health, wealth, or family. It can also come from an ego trip: you must beat this person in front of me. We mostly talk about number one, but trust me, the other two sources make up a large chunk of this.

Sar: But they don't make for a good startup origin story.

Aakrit: Yes, they don't make a good story. Capital becomes a commodity if you have a conviction from any of those things. Then you can figure out a way to raise money. We raised a seed round. We then raised a convertible note. We then raised strategic financing from Times Internet. We then raised another convertible note on top of that. We eventually sold to the Jio subsidiary of Reliance because I wasn't okay with this company failing, even if that meant it would become a part of something else. And then, as you said, post-acquisition, things worked out.

Sar: I want to talk about design constraints. Talk about messaging interface as a canvas for software experiences. A big reason people gave up on messaging context was the nature of design constraints of being beholden to five to six dominant platforms with varying levels of capabilities. There were a lot of discussions around what is best suited for websites versus apps versus chatbots. When I was deep in this market, I believed visual experiences were better for discovery-based, low-intent experiences, while chatbots were better suited for high-intent search-based experiences. That used to be my framework years ago.

Aakrit: We concluded that conversational UI was a design problem more than a machine learning problem. We would have this debate with VCs and founders all the time. We believe the NLP stack will get commoditized over time, and the product will only work through great user experience and solving the particular use case well. I think there are different categories of bots.

First is messaging only; a classic chatbot within WhatsApp or Messenger. I don't think the interface of the first category allows for most complex transactions to be completed solely on the interface, which means it can be a good starting point. You can triage, but you must go through other interfaces to finish the last few steps. That could be getting directed to a human agent or a website. You don't want to complicate it or add too many recommendations. This is meant for high-intent activities, as you mentioned. You just give the person what they want through the decision tree of the chatbot, followed by another commonly used interface to finish.

The second is voice as the primary medium. We see this in speakers and voice assistants. The category of voice is nothing but just a different form of search. It doesn't necessarily work most of the time, even today.

The third one is full-blown, two-way voice interactions, typically at a contact center. This requires the greatest amount of intelligence and technology combined with design and is the most expensive but by far the highest impact if it works.

Sar: During the peak chatbot hype mania, the prevailing narrative amongst the advocates was that bots were the new apps. They would cite increasingly noisy app stores, app fatigue amongst users, and growing attention share of messaging apps as the drivers for that. That future never played out. The conversational interface was simply too complicated and not meaningful enough improvement over apps in the West. I used to believe that chatbots had a better shot in markets like India because of hardware and mobile operating system fragmentation, expensive mobile internet, low local phone storage space, and cheap phones at the time. Chatbots felt like a low-cost and easy way to access digital services. Many of those constraints have now gone away in post-Jio India. Did you ever believe bots would become a dominant interaction paradigm and how do you asses its growth in India? Relatively speaking, India has had more success on this front because of the propensity of companies to build native Whatsapp experiences.

Aakrit: Look, at the heart of your question, lie two fundamental dynamics:

Distribution: In the US, there was no inherent messaging platform that became omnipresent, like WeChat in China or WhatsApp in India. This is why bots have succeeded here more than in the US.

Design: There are no major use cases (besides maybe customer support) proving that Conversational UI is exponentially better than the traditional GUI. And if that’s not the case, then there is very little motivation for the end customer to try to learn something new.

Sar: You have worked in this space for nearly a decade. What big bets or hypotheses have proven true so far? What has been harder than expected?

Aakrit: What’s been proven true is consumers definitely love messaging. It’s the most preferred communication channel of the 21st century. It’s async; everything can be tracked, it’s fun (with emojis), and it also makes you think before “saying” anything.

What’s been harder is making use cases beyond P2P work. Customer support has shown some legs, but it’s more a push by brands than the actual NPS going up. In India, we are seeing marketing campaigns/notifications take off in a big way on WhatsApp, but that’s not a world-changing solution - it’s an extension of SMS. I worry that WhatsApp won’t lose its reputation in the quest for revenue. Let me put it this way: I had expected more of the Conversational UI.

Sar: Discovery was a huge problem back in 2016. It doesn’t feel like sufficient progress has been made on this front. Many hoped for contextual discovery to emerge in messaging contexts to serve users what they need at the right time to drive awareness of new bots. Who did you think had the most potential at cracking this in a couple of years? Were there any promising attempts? Apple made a big deal of the iMessage App Store, but it has always been so janky. It feels like they have given up on improving it, so we are left with this annoying mess in a narrow bar above the keyboard in the app.

Aakrit: That’s correct. I always felt the consumer plays, including Haptik, have failed because of that. We all waited for the next App Store, but that never really played out. Back in the day, everyone felt Messenger would be the one to crack it, and apps like Poncho/Swell did whatever they could to figure out usage, but the underlying platform was just not sticky enough. These things work if the messaging platform is used 20-30 times daily (like WeChat and WhatsApp).

We recently launched the Haptik-powered Reliance JioMart experience inside WhatsApp. One of my former colleagues from 12 years back tweeted, “I'm so happy that this is finally happening for you a decade later.”. For the first time, I'm finding a platform on Whatsapp that genuinely solves it.

Sar: How is WhatsApp solving the discovery problem?

Aakrit: Today, through push notifications. They now enable businesses to send notifications about promotions, offers, alerts, transactions, etc., without requiring an explicit opt-in. The read rate on these is >50%, and the response rate is 50% of the opens (overall ~25%), which is game-changing if you compare it to any other push marketing channel such as SMS or email.

Eventually, this will make the biggest difference when they add a Business Directory to the main app itself. They are already testing this in Mexico, and should be available worldwide soon. Essentially, you can see all available Business Accounts in the app and discover them just like you do a contact. This will be the closest to the Apps-WeChat experience we will get and will be transformational.

Sar: How has your role as the founder evolved over your journey as an independent company and now as a part of the Reliance group?

Aakrit: Honestly, things didn’t really change after the Jio deal. They are very public that Haptik is an “investment” for them, not an acquisition. As a result, the day-to-day of our life has been pretty much like a growth stage startup at Series C/D, and we also get measured that way. And from that perspective, my role has changed significantly like any other founder of a growth-stage company. I am less involved in solving problems and more in managing people and capital. The Reliance and Jio brands bring some credibility, which becomes important in enterprise sales & partnerships.

Sar: What can you tell us about the culture of Jio as a company?

Aakrit: Jio is a unique place. I have never seen a culture like theirs. They essentially operate like a garage startup behind the veils of a corporation. Their speed of execution and decision-making is second to none. They are extremely bold and not afraid to make decisions worth billions over a few seconds. But the biggest misconception is that they are not “fair” or “hard asses” to deal with. I have found that to be completely the opposite. They are super straightforward, transparent, and walk the talk. If you perform, you will succeed. They also firmly believe that it will never succeed unless all parties make money in a given deal.

Sar: What have you learned about maintaining the balance between sticking to a strategy and not avoiding opportunism?

Aakrit: Yeah, good question. I’ve learned that the way to achieve this balance is to have it within the leadership. Most Founder CEOs are opportunistic problem solver types. They love working on unbounded problem statements and solving them. They can be random at times. I am a little bit like that.

So, to balance that, you need the other cofounders or execs to be the ones who are less creative and more operational. It could be a COO or CTO, or even CFO.

It could be reversed that a particular founder is more operational, but that’s rare. The so-called mavericks have created the most outsized innovative companies, and you need to continue to have that culture of startup chaos no matter how large you get.

Sar: What is underrated about Bombay as a tech hub? What is overrated about Bangalore? What tier-two cities do you see a surprisingly large number of promising angel pitches from?

Aakrit: Right, I’ve been spending a lot of time on Bombay vs. the rest. Bombay has a branding problem as a tech place. By the last count, there are at least 20 unicorns that are entirely headquartered here. The cost of living debate is not even relevant today, where Bangalore salaries are higher than anywhere else.

I have always felt Bombay has a culture of building real businesses, ones that are focused on the first principles. You get carried away in the funding and valuation echo chamber in Bangalore. Most people living in Bangalore would agree that Bombay is a nice city to live in. I can’t think of any surprising Tier 2 cities these days, but there are many remote-first teams whose Founders are sitting out of nowhere.

Sar: You have been making angel investments in India for years. What kind of problems do you wish you saw more Indian founders go after?

Aakrit: I tweeted about this recently. Most founders in India, including me, have been guilty of going after capital and valuation. The media has glorified this story so much that we all think that’s the end game. I can’t think of more than 2-3 founders in the Indian tech startup ecosystem who genuinely solve problems they love. I think there is a HUGE opportunity to go after India-first problems such as b2b supply, energy usage, infrastructure, etc. Look at spaces where there are existing large markets already and how technology can provide a better way of solving problems within it.

Problems such as :

Digitization of the supply chain across commodities such as chemicals & textiles
Electric vehicles, battery tech, charging infra
Climate: carbon removal, solar energy
Banking the underserved
Healthtech

Unfortunately, most of these business models don’t fit the classic startup VC world. They take a long time before they reach an inflection point. You can’t become a unicorn on the back of showing DAU/MAU and a dream.

Couple of companies I like: myUpchar, DeHaat, Fashinza

Recent conversations :

Demystifying developer platforms with Ceci Stallsmith, Cofounder of Calyx

A gutsy rollercoaster with Todd Saunders, CEO of Broadlume

The business of internet building with Joseph Cohen, CEO of Univer.se

Making demos interactive and fun with Caroline Clark, CEO of Arcade

Brewing coffee & growth with Lauryn Isford, Head of Growth at Airtable

Dialect-based content in India with Vinay Singhal, CEO of STAGE

Bringing software to public interest with Jimmy Chen, CEO of Propel

Demystifying product led sales with Alexa Grabell, CEO of Pocus

Coming of age of fintech in India with Priyanka Kanwar, Cofounder of Falcon

Transforming farming with Jayce Hafner, CEO of FarmRaise

AngelList Stack is for startups that want faster fundraising, cleaner cap tables, and high interest banking, all in one place. Scaling companies such as Abound, Harness Wealth and Syndicate migrated from other vendors in less than a week with zero legal fees to gain an unfair advantage in managing their back-office. Learn more by signing up here.

Scatter Brain

A decade's pursuit in a category filled with…