Swapan Rajdev

The through-line

Across nearly a decade of pivots, Rajdev keeps returning to a single founding thesis that he says never changed even as the company's shape did: the experience between a user and a brand "didn't feel the most smoothest or the best" [1], and the rise of digital conversations was the trend that could fix it. He frames that customer experience problem as end to end rather than as customer support alone, covering how a user discovers a brand, finds the products and services that suit their needs, gets post-sales help, and later recalls the brand at all [1]. The other half of the thesis came from watching WhatsApp gain traction around 2011 and BlackBerry Messenger before it, and concluding that "we as human beings are conversational in nature" while the way people communicate is becoming far more digital [1]. Haptik, in his telling, is the attempt to "take this trend that we're seeing of digital conversations and these digital mediums take this problem statement of customer experience and marry the two" [1].

What shifted was everything else. The company pitched businesses first and found "none of them cared" [1], moved into survival mode and built a B2C everyday personal assistant app covering reminders, grocery shopping, movie tickets, cabs, flights and restaurants [1], then returned to enterprises around 2017 when the market turned and brands began asking Haptik to build bots for them, at which point the B2C arm was shut down and the company became a pure B2B SaaS business [1].

On conviction as the only thing that gets you through

Rajdev is blunt that money is the wrong motivation. He admits there were times he thought "I'm doing some of this because I think I can make a lot of wealth", and his verdict is that "whenever I thought through that lens you never succeed", because at some point you conclude "this is not worth it" [1]. What replaces it is belief, and specifically belief in a future state rather than the present one: not the business model or the product as it stands today, but "a world that you imagine three five years down the line" [1]. He says Haptik considered shutting down multiple times, both very early and very late, and that even with a couple of million users on the consumer app the existential question persists: "will anyone's life change if I shut down today" [1]. The company ran out of money twice; he and his co-founder Akrit put in what little savings they had left and borrowed money from a friend to keep it alive [1]. The reason he now describes the outcome as validation is that the imagined world arrived: "luckily we are sitting there where everyone is using chat bots" [1].

He is equally direct about the structural support that makes conviction survivable. Having a co-founder is "so much better than being a single founder" [1], because the buck stops with the founder, and while success is shared, "the failures are definitely yours" [1]. With even ten or twenty people working on a vision only the founders hold conviction in, the doubts turn inward: "am I leading people down the wrong way am I chasing a pipe dream that will never come true" [1]. Two people going through the same downturn can take turns absorbing it, and where there is no co-founder he suggests a panel of mentors or advisors to talk to instead [1].

On creating a category before the market exists

Before the space boomed, Rajdev says chatbots, IVAs and virtual assistants "didn't exist then and we were one of the first few people really pushing for this" [1]. The work at that stage was education rather than selling: convincing people not only that the technology was possible but that it was genuinely a good experience [1]. He credits the survival of that period to "the right people took the bets on us" [1]. The market flipped in 2017 when Facebook, Amazon, Microsoft and Google all entered, enterprise behaviour changed, and brands began approaching Haptik directly after seeing the consumer app [1]. The pivot back to B2B was possible because the consumer product had already forced the technology into existence: Haptik had been building bots for itself, so it could redeploy that stack on customers' channels, and from there revenue flowed and the business model added up [1].

On what stays the same between consumer and enterprise products

Rajdev separates the fundamentals of product management, which he considers constant, from the mechanics, which are not. The two universals are distribution and messaging. Whatever you build, you have to think about how you will get distribution, and then about positioning, which he reduces to a single test: "it's always been expectations versus reality" [1]. In consumer that meant the website, the ads, the Play Store and App Store pages all setting a promise, and the product either meeting it when the user arrives or not [1]. Get the positioning right for what the user actually wants and the outcomes follow [1].

Design carries that same burden of expectation setting. He returns to the personal assistant days to make the point: even though Haptik had stated it was a personal assistant app and defined what it did, users would still ask things like when a film star's birthday was, because "we left the canvas blank" [1]. A blank canvas invites anything, which is why he says "design plays a big role" [1] in a conversational product.

On the mechanics that differ between B2C and B2B

The first difference is the feedback loop. In consumer you cannot easily talk to users, so feedback has to be "fully analytics driven", but the loops are short: you ship something, watch usage, and quickly know whether it works [1]. In enterprise you can talk to customers constantly because you sell to them directly, and the quality of feedback that comes back is high, but validating anything takes longer because sales cycles are longer [1]. The second is tolerance for error. Consumers are "very unforgiving" and a bad first journey means "you've lost the guy" with little chance of recovery, whereas enterprise relationships can be maintained and repaired, and a customer with whom you have built a good relationship will give you another shot [1]. He notes the irony that B2B products are the more complex and confusing of the two, yet the ones where you get more than one attempt [1].

The third difference is who does the positioning. In B2B, distribution is not purely digital, so salespeople have to be enabled to convey "the true power of the platform" [1]; the founder's job is making sure they understand the product well enough to represent it. That drives a whole different process set around go to market: positioning, sales enablement, regular website updates, feature pages and one-pagers [1]. On the consumer side the equivalent work is user onboarding journeys, discovery and pop-ups that catch attention [1]. Different processes to reach adoption, same underlying fundamentals [1].

On the buyer and the user in enterprise sales

Rajdev acknowledges the enterprise puzzle of the user and the buyer being different people, and that a decision can involve ten different people including influencers, describing it as a genuinely interesting tension between building for the user and building for the business case [1].

Takeaways

• The founding thesis has stayed constant through every pivot: marry the rise of digital conversations to the end-to-end customer experience problem, covering discovery, product fit, post-sales help and brand recall, not just support [1].
• Motivation built on wealth fails; the durable version is conviction in "a world that you imagine three five years down the line" rather than in today's product or business model [1].
• Haptik ran out of money twice, with the founders putting in personal savings and borrowing from a friend, and considered shutting down multiple times even after reaching a couple of million consumer users [1].
• A co-founder matters because failure lands on the founder alone; if you do not have one, build a panel of mentors or advisors to talk to [1].
• The B2C personal assistant app funded the technology that made the B2B pivot possible, since Haptik had already built bots for itself and could redeploy them on customers' channels [1].
• Distribution and expectation versus reality in messaging are the fundamentals that hold for any product; leaving "the canvas blank" in a conversational interface invites requests you never promised to handle [1].
• Consumer feedback is analytics-driven with fast loops and unforgiving users; enterprise feedback is conversational and rich but slow to validate, with more room to recover a relationship [1].