Shantanu Sathe

India has abundant expertise across every domain - logistics operators, subject matter experts, local service providers, etc. What's scarce isn't capability. It's the systems that coordinate this capability at scale. Most companies see this fragmentation and default to vertical integration: build everything in-house, own all capabilities, control every aspect of delivery. This works if you have unlimited capital and time. Most companies have neither. But there's an alternative approach that we've noticed two of our portcos execute remarkably well, in completely different industries - 1. When Meesho looked at India's logistics, and they saw thousands of fragmented local operators. Instead of building warehouses and delivery fleets, they built Valmo - an orchestration platform coordinating existing partners. Individual pilots with smartphones earn sustainable incomes, small hub operators build viable businesses, and Meesho gets coverage across 15,000+ pin codes at 12% lower cost than traditional third-party-logistics. 2. Entri saw similar fragmentation in education. India has abundant teaching expertise already running offline/online programs. Entri partners with them for content while owning demand generation, platform infra, quality monitoring, and placements. This allows them to launch new categories in weeks and scale across languages without the capital intensity of building in-house teams for every vertical. If I had to pull a common thread, it's that in fragmented markets like India, the orchestration layer often matters more than asset ownership. Coordinate what exists rather than rebuilding from scratch.

31

1 Comment

I constantly oscillate between sharing in-progress work and completed work. This means I'm long overdue to share some exciting updates on Wonderment Delivery Promise, our pre-purchase ETAs product. Wonderment now supports Rules for transit estimations and fulfillment estimations to allow you to account for per-product ETA customizations (like embroidery and engraving), warehouse delays, or other events that would impact coming estimates. Here's a quick demo of how you can use them for communicating pre-orders, backorders, and delays: https://lnkd.in/eb88vmcu

36

1 Comment

Praveen Neppalli Naga, Uber's CTO, was kind enough to join us a few weeks back (continued from what my Partner, Ganesh Srinivasan previously shared) to walk us through his philosophy on building operationally complex AI systems. And as always, it all leads back to fundamentals. Even the best AI model can’t outrun a messy data stack. As Praveen put it, scalability isn’t a function of model sophistication; it’s a function of data quality, flow, and freshness. Any cracks in that foundation will eventually become the true bottleneck to innovation. This was just one snippet among many learnings, and it was such a pleasure speaking with Praveen. Thank you again to him and the Uber team for the warm welcome at their offices. I'll be keeping this lesson top of mind as Ganesh and I head to #NeurIPS2025 this week. We're excited to see what innovation is around the corner. Please reach out if you’d like to chat. We’d love to meet up on the convention floor.

77

3 Comments

When I started as a #PM, a large part of the role was writing #PRDs and iterating with stakeholders before handing off to engineering. That process is rapidly changing and PRD's as we know it are dying. Here’s what will replace it. 📋 FROM: Traditional PRD process 📋 ■ Spending hours interviewing customers and manually synthesizing insights ■ Drafts that miss standard language and cross-stakeholder expectations ■ Low-fidelity or no visuals in early drafts ■ Lengthy, repetitive review cycles with multiple stakeholders ■ Manual updates after every meeting ■ Manual breakdown of PRDs into features and user stories ■ Manual creation of QA test cases and UAT scripts ■ Static, outdated documentation post-launch ⚡ TO: GenAI-enabled PRD process ⚡ ■ AI-augmented synthesis with interrogation and summarization capabilities ■ GPT-powered co-creation that aligns across stakeholders and domains ■ Instant generation of interactive prototypes for each feature or improvement ■ Real-time prompt refinement during working sessions to shape ideas ■ Automated artifact updates from meeting transcripts and decisions ■ Auto-generation of features, user stories, and acceptance criteria ■ Automated creation of QA test cases and UAT scripts ■ Continuously evolving, real-time documentation synced with changes 🧠 What it means for PMs 🧠 1️⃣ You’re no longer measured by specs, but by how fast you validate and ship ideas 2️⃣ Documentation isn’t dead, it’s now a dynamic, AI-supported asset 3️⃣ The best PMs don’t “own the doc”, they orchestrate the system Shai Dinnar Diana Cordova Shivani Rathi Emily Gao Dane Greisiger Justine Wang Kushal T. Anne Schmitt Bradley Antcliff Dimitrios Lippe

93

8 Comments

Would love to hear thoughts from folks in logistics, freight, and transport tech: Tracking market vehicles (spot-hired, non-owned trucks) in India continues to be a huge challenge. SIM-based tracking has been one approach, but in reality: 1. Getting driver consent is a task in itself 2. It needs handholding every time: OTPs, setup, approvals 3. Drivers often switch off phones or SIMs mid-trip 4. Uptime and reliability remain inconsistent Driver apps are widely used in the US and other countries, but adoption in India is low. Many drivers either don't use smartphones regularly or actively disable location sharing. Yet, this segment is critical, market vehicles move a massive volume of freight in India. So the question is: 👉 What will it really take to track market vehicles in India with better uptime, accuracy, and less friction? If you're solving this or struggling with it would love to hear your experiences. #LogisticsIndia #FleetTracking #TransportTech #SupplyChainVisibility #LogisticsTech #RealTimeVisibility #MarketVehicles #TripTracking

160

27 Comments

When we think of self-driving cars, names like Tesla or Waymo usually come to mind. But here is something most people don’t know about India's autonomous driving technology. Even India has its own AI-powered pioneers quietly building systems that are uniquely designed for our unpredictable and high-density traffic. 🔹 Minus Zero – A Bengaluru-based startup that created the world’s first camera-only 'Level 5' autonomy system. Their 'zPod prototype' proved that cars can navigate chaotic Indian roads, without depending on expensive LiDAR. 🔹 Swaayatt Robots (स्वायत्त रोबोट्स) – From Bhopal, they’ve trained Mahindra Boleros to autonomously handle toll plazas, bi-directional roads and mixed traffic scenarios that global players usually struggle with (remember the viral video from last year). 🔹 Flux Auto – Backed by Y Combinator, they are not just working on cars but also 'retrofitting' trucks and tractors for mining, logistics and agriculture (areas that could transform India’s supply chains). Even big players are moving fast: - Tata Motors has filed multiple patents in 'autonomy' - Mahindra Group in working with its Tech Mahindra arm to explore advanced driver-assist systems. This ecosystem shows India isn’t just catching up. It is shaping affordable and adaptable autonomous mobility for the world. 💡 How do you see the future shaping up for India in AV space? #AI #AutonomousVehicles #IndiaTech #Mobility #Innovation #TheCuriousMinutes

2

Some of the most meaningful engineering work happens in unexpected places. Rivian’s haptic steering wheel is a great example of a this—a meaningful innovation we didn't realize we needed! But one use and it's obvious this is a leap forward in usability. Vidya Rajagopalan, Rivian's SVP of Electrical Hardware, and her team blended hardware and software to build an adaptive system that gives drivers control and tactile feedback without distraction. It’s elegant engineering and an experience that feels immediately intuitive. Follow along for more: https://lnkd.in/gJUq4xyW

923

12 Comments

India may be the land of spices… …but the spiciest opinion in product management comes from Marty Cagan — the man behind Inspired, Empowered, and Transformed. In this week of Lenny & Learn, I went down the path of questioning my existence- all thanks to Marty. If no one has called him the father of modern product management, it’s a shame, as he is definitely like a father tough, brutally honest… and almost always right Marty answers a very important question that comes across the minds of both tech and non-tech world- Who really is a product manager and what do they do? The best product led companies work on four principles - Usable, Feasible, Valuable & Viable. Now, usable & feasible are what designers and developers respectively are responsible for, but the valuable i.e. Is the company going to make money and the viable i.e do customers need it, are the top most if not only responsibilities of product managers. Product managers need to think about problems, but they need to think more about solutions. Solutions that are not a list of features but those that create value. They focus on outcomes and not on output. To loosely explain Output vs Outcome- Output is the count For ex: number of features, number of campaigns, number of widgets etc and Outcome are the changes that result from the outputs i.e. Increased user engagement because of a feature, Increased sales because of a campaign etc. Understanding this concept, actually made me understand what makes someone a great product manager: They really know about the users and customers (Meet X customers before making a decision) They are an expert in data of how product is really used, sales analytics, marketing analytics They know the industry, the market in which business operates in and more They know the different parts of business- how is it marketed? How is it paid for? What are the complaints, regulatory requirements? The purpose of knowing and doing these is to innovate, make a valuable and viable product, but most importantly be able to approve or improve a prototype in the shortest period of time. The ability to innovate at the optimal speed is what differentiates a great company from the rest, and product managers are at the centre of that world. If I have to summarise, all Marty’s sessions, SVPG’s blogs- It would be that Product managers are solving What is just now possible For real users or customers In the ways that helps the business. And to do this the best and the only way is for them to have: Unencumbered access to users and customers Unencumbered access  to engineers Unencumbered access to stakeholders These sessions shook me (I am sure it did Lenny too), but sometimes you need tough love to do your best work. P.S: I would recommend Teresa Torres book- Continuous Discovery Habits, which comes highly recommended by Mighty Marty.

5

Harsh but the truth. Domain knowledge may help you win in the short term, true product skills help you win in the long-term.

6

This is a great, provocative take by Philippe De Ridder on how AI is reshaping product development, shifting the bottleneck from engineering costs to user attention. The discussion that follows is equally insightful! While many agree that attention is the new battleground, a significant number of commenters emphasize that the core "Build-Measure-Learn" principles of Lean Startup are more relevant than ever. AI simply allows us to iterate and learn faster, making higher-fidelity prototypes and testing real user behavior more efficiently, rather than skipping essential validation. The consensus leans towards AI accelerating Lean, not making it obsolete, with the crucial need for deep customer understanding still paramount to avoid building "the wrong thing faster."

2