Aug 14, 2026 · 1h 7m · neon-show
ChatGPT Killed My $50M ARR Business Overnight | Swapnil Jain Observe Al
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Observe AI co-founder and CEO Swapnil Jain details how the rise of generative AI forced the company to dismantle its successful $50M ARR human-agent SaaS business and pivot entirely to autonomous AI agents. He shares critical lessons on cross-border enterprise go-to-market, wartime leadership, and navigating technological disruption to build an enduring enterprise platform.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Siddhartha as informed peer 3.9 Guest teaching 4.1 Guest disagreement 2.3 Siddhartha pushing back 2.6
speaking balance: gold is Siddhartha, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 45:04 Challenging hyped AI startups and contracted ARR
Swapnil forcefully rejects the host's premise that frontrunners like Sierra and Decagon have secured the market, calling their ARR unproven contracted revenue and Silicon Valley fluff.
Hardest push from Siddhartha ▶ 32:36 Halting the guest to challenge a pivot during strong growthSiddhartha directly interrupts Swapnil to question why Observe initiated a painful internal disruption when they were still scaling rapidly without adverse external signals.
Biggest teaching moment ▶ 1:00:00 Reframing point-solution voice bots as commodity featuresSwapnil educates the host on enterprise purchasing reality, demonstrating why hundreds of standalone voice AI startups will fail without end-to-end operational toolsets.
Siddhartha holds their own ▶ 27:22 Host counters underdog narrative with global viewership dataWhen discussing feelings of inferiority building from India, Siddhartha demonstrates his own market leadership by citing his podcast's 120M views outpacing Y Combinator and A16Z.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Siddhartha as informed peer | Guest teaching | Guest disagreement | Siddhartha pushing back | Why |
|---|---|---|---|---|---|---|
| The ChatGPT Shockwave and Re-evaluating Observe AI | 4 | 3 | 2 | 3 | The host sets a blunt opening premise that ChatGPT made Observe AI's initial quality-assurance business obsolete. The guest agrees with the framing and elaborates on how LLMs forced a foundational pivot to Observe 2.0. | |
| Disrupting a $50M Business for Observe 2.0 | 4 | 4 | 2 | 3 | The host notes Observe was a top Indian-origin SaaS company that suddenly went quiet after raising at an $850M valuation. The guest reveals they consciously sacrificed short-term growth in a $50M ARR business to re-engineer as an agentic platform. | |
| Creating the Customer Service AI Market | 4 | 3 | 1 | 2 | The host inquires why Observe attracted massive venture backing early on. The guest details creating the market for AI in customer support before tracing his transition from Twitter in SF back to India. | |
| Assembling Co-founders and Focusing on Support | 3 | 3 | 1 | 2 | The host asks how the founding team converged and shifted from sales enablement to customer support. The guest explains his co-founder's domain knowledge of call center hubs in Manila. | |
| The Costly Mistake of Selling to BPOs | 3 | 5 | 3 | 2 | The guest explains why selling to BPOs was a fundamental error because BPOs are execution vendors rather than software buyers. He also strongly warns against selling software in India, sharing a story of an enterprise client defaulting on a $500K contract. | |
| Launching US GTM, Y Combinator, and Overcoming Insecurities | 5 | 3 | 1 | 2 | The host probes the psychological transition of moving to the Bay Area and feeling like an underdog. The host demonstrates his own standing by citing his show's view counts against top Silicon Valley podcasts. | |
| Scaling Observe 1.0: Seed to Series C Milestones | 4 | 4 | 1 | 2 | The host prompts for the chronological financial milestones of Observe 1.0. The guest provides detailed revenue figures from $1M ARR at seed to $15M ARR at their SoftBank-led Series C. | |
| The Shift Toward Full AI Agent Autonomy | 4 | 5 | 3 | 4 | The host halts the guest to question why they disrupted their core business when metrics were still scaling well without external shocks. The guest explains that keeping the SaaS model would have capped them at a mediocre outcome as autonomous agents take over. | |
| Co-founder Transitions and Leadership Evolution | 4 | 3 | 1 | 2 | The host asks about the emotional and operational impact of original co-founders departing. The guest details promoting their chief scientist to co-founder and treating the executive team as equity partners. | |
| The Future of Customer Support: Headcount vs Volume | 5 | 5 | 3 | 4 | The host and guest discuss whether AI support agents will reduce human headcount or if overall ticket volume will expand. The host pushes back by pointing out widespread layoffs in software engineering teams. | |
| Questioning AI Market Hype and Startup Sustainability | 4 | 6 | 7 | 4 | The host asks whether heavily funded AI startups like Sierra and Decagon will dominate. The guest forcefully dismisses the hype, arguing their ARR is contracted rather than deployed and calling current valuations Silicon Valley fluff. | |
| Reverting to Wartime Startup Mode | 3 | 3 | 1 | 2 | The host asks what the guest focuses on day-to-day. The guest describes running the company in hands-on wartime founder mode across sales, engineering, and cultural resetting. | |
| Mastering Go-to-Market as a Technical CEO | 4 | 4 | 2 | 2 | The host asks how a technical founder mastered go-to-market execution after his GTM co-founder left. The guest explains that non-infra CEOs must lead sales and marketing personally through empirical trial and error. | |
| Managing Team Turnover and Cultural Reset | 4 | 4 | 1 | 2 | The host asks about the difficulty of enforcing a cultural overhaul. The guest candidly shares the friction of employee departures, self-reflecting on his strengths in resilience and flaws in decisiveness. | |
| Decadal Commitment and Scaling via GSI Partnerships | 4 | 4 | 2 | 3 | The host questions why the founder didn't start fresh with a new entity after ten years. The guest affirms his decadal commitment to Observe and outlines their shift toward Global System Integrator partnerships. | |
| Voice AI Commodity vs Enterprise Lifecycle Platform | 4 | 6 | 5 | 3 | The host notes that voice AI is saturated with hundreds of startups. The guest educates the host on why standalone voice bots are commodities, arguing that only unified platforms managing the full agent lifecycle will survive. | |
| Core Advice for Enterprise AI Founders | 4 | 5 | 3 | 2 | The host asks for distilled lessons for enterprise AI founders. The guest advises targeting massive US markets, avoiding price wars, and actively redefining the customer's evaluation criteria. |
Statements from this episode (25)
Observe AI is growing 5x year-over-year after pivoting to AI agents
“And that's sort of the new observe two point or that we are seeing where we are going five X year on year going forward.”
Observe AI reached over $50M in revenue before pivoting to AI agents
“So, I mean, I can't share the specific numbers, right? But it was not of fifty million dollars, right? So we reached a pretty significant scale selling our previous generation technology, previous generation software”
Jain: Observe AI created the AI customer service market
“So in many ways, I think of we created this market. We created this awareness that AI in customer service is legit. It's going to happen. And then everyone followed”
Jain: Customer service AI startups like Sierra emerged post-ChatGPT
“And then chat GPT happened and then everyone realized actually customer service is going to be the biggest market for AI. And they've been on the journey since then. You know, I think in customer experience in customer service, Sierra, Decagon, everything happ…”
Jain: Voice AI for customer service has a larger TAM than sales
“As I spend more, more time in there, I very quickly realized that it's not that big of a TAM and that world is ending very, very fast. On the other side, Sharath had spent a lot of time in Philippines in his previous job, so he convinced me that no, customer s…”
Jain: Selling software directly to BPOs fails because end-clients control tech stacks
“The end client is a decision maker. They decide what technology stack will be used, and BPO is just doing that. So trying to sell to a BPO doesn't make sense. You have to go to the end enterprise.”
Jain: Indian SaaS startups cannot enter the US by just hiring reps
“And a lot of times what founders end up doing is they say, okay, it's so easy. I just hired two sales rep in the US. And I'm like, never, it's never, ever going to work because it's not just about your sales. Your entire company is meant and built For Indian n…”
Jain: Observe AI's only Indian customer ghosted a $500K contract
“We've ever, we only ever, ever sold to one Indian customer, and that's because it came from an investor, a mutual investor, a great brand in India public, they recently went public, so all, all amazing things. They were a customer very large customer, got it l…”
Jain: Feeling small in Silicon Valley drives founders to be more ambitious
“I don't think there will ever be a time that the place will make you feel big. Because there's so many amazing things happening in Silicon Valley and Bay Area that no matter what you end up doing, right, you will always feel small, which is, Part of, which is …”
Jain: Observe AI stopped selling its legacy product to force a pivot
“We almost like, yeah, we stopped selling. We incentivize all of our sellers and marketing team to just focus on newer products. We still get demand for our older product, but we've also transformed that to agents. So now we are selling agents.”
Jain: Customer service is a $400 billion industry employing 20 million agents
“Right now roughly there is like twenty million customer service agents, roughly four hundred billion dollars of spend, both on people and software included.”
Jain: AI will handle over 50% of customer service within 3-5 years
“So on one side, I believe that if you draw it out three to five years, more than 50% of the customer service interactions will be fully handled by AI very comfortably.”
Jain: No Observe AI enterprise customers have reduced human headcount yet
“I've not seen reduction in customer service reps by any of our customers. And we have customers who have tens of thousands of agents. So we would be the first one to see the impact and the reality is nothing yet.”
Jain: High probability customer service AI startup Sierra gets killed
“There's a very high chance Sierra will get killed.”
Jain: Sierra and Decagon's reported ARR is contracted, not deployed at scale
“And the reality of a lot of, in full reality, a lot of the so-called ARR that you see with some of these companies that you mentioned is actually contracted ARR. So none of this is ever actually deployed in production at scale. Find me one real customer from a…”
Jain: I am running Observe AI in a wartime founder mode
“I'm almost running Observe and because it's Observe two point O in like a hard founder and a wartime mode, and also I would almost say like a seed to series A company. I'm not running this like a 250 people company or a series C mature scaled company.”
Jain: Enterprise culture and systems built before 2022 are completely obsolete
“You have to understand anything that was built before 2022, 2023, both from, not from a product perspective, culture, processes, systems, org, doesn't matter.”
Jain: App software CEOs must master sales, marketing, product, and engineering
“Unless you are a deep infrastructure company where the CEO could be very, very technical, In an application stack company, the CEO needs to be a great marketer, great seller, great visionary, great product person, know about engineering, know about everything.…”
Jain: Observe AI conducted layoffs and lost talent during its pivot
“As I said, as part of this process, right, we also, we had to let go quite a bit of people. But also, you know, the downside of this is a lot of good people also leave.”
Jain: Observe AI spent 18 months replatforming into autonomous AI agents
“Now everyone is 2.2. We've been on the journey for a bit now. Right. So everybody is now observe 2.2. We took the older product, right. That we built completely replatformized it. It's now all agents and whatever new builds all agents anyway, sellers are all a…”
Jain: I underestimated LLMs and should have pivoted a year earlier
“I feel like I was a little too late in the two point of transformation. Maybe we should have done a year earlier. So I could have done this sooner. I think I, Underestimated the power of LLMs and the power of where, where the very eyes going in general.”
Jain: Standalone voice AI is a commodity that won't make money
“Voice AI is commodity. That's not the future. No one is going to make money in the long run on just being a standalone, simple vanilla voice AI company. You have to build enterprise grade deep platforms.”
Jain: Indian SaaS founders must sell to the US from day one
“For a lot of India founders, I will say don't sell in India. U.S. Is the biggest market for almost every industry. Go and start selling from here. Do not even think about starting from India. You will get stuck. You won't be able to make that jump. A lot of fo…”
Jain: US software startups cannot win on price, only on product
“Third is to compete in global markets and especially U.S. Market. You can't win on Price or you know any other differentiation other than product. You have to be a phenomenal, phenomenal product to compete with companies, because every industry you pick, every…”
Jain: Enterprise AI startups must redefine categories, not fight feature wars
“You have to be categorically differentiated. So when I go into a deal, I don't say my voice AI is better. I challenge the customer and say, why are you just doing voice AI? You should do voice AI and you should do companion agent and you should think more broa…”