Hrittik Roy

Last month, I was interviewed by Broadcasting Corporation of Oyo State (BCOS) where I spoke about how social media influences human behavior, and the role AI plays in that. I shared how AI algorithms could suggest content or network connection that can either help or harm, depending on the user. While some people are able to filter what they see and engage with online, many have unknowingly been influenced sometimes negatively. That’s why I believe responsible AI is not optional. It’s something we all need to take seriously, whether as developers, users, or researchers. This calls for more work around AI safety and ethics, especially in contexts like ours where awareness is still growing. A 1-minute extract of the video is here, https://lnkd.in/dBwadpGx #AISafety #EthicalAI #SocialMediaImpact #DigitalWellbeing #TechForGood #ResponsibleAI #BCOSInterview

72

4 Comments

I asked ChatGPT the personality type of Shiva (the Indian God). Recently, I asked ChatGPT what the possible MBTI type of Shiva could be. Chat readily answered INFJ - a supposed personality type that is rare, and therefore, Shiva must be it. Now, how many times do we trust ChatGPT? Many times. Most times. We think, since it is taking in information from all over the internet, it must be doing a great job at thinking critically. But no - only somebody who has delved deep into a certain domain will be able to verify whether ChatGPT is actually thinking critically or not. In this case, I could really point it out. I questioned Chat about how Shiva is a certain type when he actually uses some other traits dominantly. That’s when Chat started rectifying itself accordingly. It also mentioned how it picked up from oversimplified explanations on the internet. But there are critical ones too. There’s a lot of in-depth, analytical information out there - and if it picks that up, and truly uses intuition, it will start seeing the connections and coming up with critically thought-out answers. Just goes to show - most human beings, and not even AI, are really thinking critically. Our AI is picking up on popular data, just like most of humanity does. What does this say about AI's intelligence? How intelligent is artificial intelligence? Maybe it’s at a working level of intelligence right now - but is it truly intuitively intelligent? Will it become so soon?

160

32 Comments

🚀 A Bengaluru startup just made history. Lyzr AI raised $8 million in Series A but here’s the twist: The entire round was managed by an AI agent named Sam. 🤯 Sam handled everything investor outreach, due diligence, and even Q&A. No missed details. No human fatigue. Just precision. Founded by Siva Surendira, Anirudh Narayan, and Jithin George, Lyzr AI is building what they call Organizational General Intelligence (OGI) AI agents that can autonomously run HR, Finance, Ops, and Customer Success. What began as an internal experiment soon turned into a full-blown fundraise completed in just 3 months. This isn’t your typical funding story. It’s a glimpse into the future where AI doesn’t just assist founders, it represents them. #AI #Automation #LyzrAI

194

12 Comments

We talk to at minimum 10-15 software teams in India a week, and most are willing to spend money on AI agents for the software development. A founder of an early stage company I met said they have given $50/month budget to every developer in their team. Until this AI wave, most engineering teams used free/community version of tools like VSCode/JetBrain/GitLab Community etc. This is the first time engineering teams are willing to pay for AI tools for their engineering team, the same price that global engineering teams are paying. There are roughly 5m developers in India, second to China and bigger than the US. If on average, engineering teams spend $500/dev/year (about 42K INR), that’s USD 2.5B opportunity. This might be the single biggest SaaS opportunity in the Dev tool space, and assuming China is hard to crack, could be the largest market for the startups/bigco in the Dev AI space.

135

8 Comments

“Speed gets you downloads. Stickiness gets you loyalty. 💡 In my capstone project at Airtribe, I reimagined how Snabbit could turn one-time users into lifelong evangelists. 🚀 ” When I was handed Snabbit (India’s first 10-minute home services app) as my capstone casestudy, I knew I was stepping into a world full of complexity, competition, and opportunity. The challenge? Snabbit had nailed speed but struggled with stickiness - users loved the idea, but only 15% returned for 2+ weekly services. So, I rolled up my sleeves and went deep: 🔍 Mapped the Jobs-to-be-Done across functional, emotional, and social needs. 📊 Conducted market research, user surveys, and competitor benchmarking. 😬 Identified pain points: inconsistent service quality, low trust, and lack of habit-forming features. The turning point was reframing the problem: Snabbit wasn’t just fighting for bookings, it was fighting for loyalty. That insight led me to design 5 Retention Weapons: 🤝 Preferred Providers → Same expert, every time. 🎊 Festival Bundles → Tapping into cultural rituals for recurring demand. ⭐ Expert Profiles → Trust before the service even begins. ⚡ Blitz Mode → 15–30 min panic cleans for last-minute needs. 🤖 AI Concierge Chatbot → Conversational, frictionless bookings. Each idea wasn’t just a feature - it was a growth engine designed to turn one-time users into repeat evangelists. From user journeys to GTM rollouts, OKRs to North Star metrics, this project stretched every product muscle I had - and helped me think less like a student and more like a product leader in action. Grateful to my mentors (Rahul Chopra, Saurabh Misal, Anmol Rattan, Karan Sethi, Kumar Utsav), the Airtribe team (Dhaval Trivedi, Shubham Chanchawat, Dhriti Singla) and my cohort buddies who pushed me to connect the dots, question assumptions, and build with both heart and rigor. 🙏 This wasn’t just a capstone project, It was a crash course in building a product that sticks. #ProductManagement #CaseStudy #CapstoneProject #UserRetention #GrowthStrategy #CareerComeback #OpenToOpportunities #Snabbit #UserRetention

32

10 Comments