Ankush Agarwal

Ankush AgarwalJul 30, 2026

36 hours can teach you more than 36 days sometimes. Last week, I took part in my first hackathon in anchors. For 36 hours, we were in a loop—writing PRDs, building, getting our ideas grilled, grilling everyone else's, changing directions midway, and somehow still finding the energy to keep going. By the end of it, I shipped 2 PRDs with working frontend prototypes. Were they perfect? Nope. But that's kind of the point. You stop trying to get everything right on the first attempt and start making things better with every conversation, every review, and every iteration. The biggest takeaway wasn't even the product. It was watching a room full of people care enough to challenge every idea—not to prove someone wrong, but to make the product better. Somewhere between the midnight food runs, random jokes, and heated debates over tiny product decisions, 36 hours flew by. I walked in expecting to build features. I walked out with a much better appreciation for how good products actually get built. Can't wait for the next one. Onwards. 🚀 #Hackathon #ProductManagement #BuildingInPublic #StartupLife #Learning

Ankush AgarwalAug 11, 2026

"Automate everything" is the pitch. Who owns the automation is the actual fight. For years Make (formerly Integromat) was just the prettier Zapier, and n8n was its scrappier open-source cousin. That framing doesn't hold anymore. Look at where each company placed its bet and you're looking at two different theories of how the AI-agent era gets won. They didn't really compete on nodes. They competed on who owns the trust layer. Make stayed closed and fully hosted. One polished stack, one vendor, no visibility into what's running underneath. n8n went the other way — source-available, self-hostable, you can drop into raw code at any node. Same category, opposite answer to the question that actually matters now: Will you let a black box run workflows that touch your data and call your LLMs? It's a product decision wearing a licensing decision's clothes. That single choice is why n8n's interface stayed thin and utilitarian while Make invested in making the canvas genuinely nice to watch. And the market already picked a side. Make sits inside Celonis with no standalone valuation of its own — a feature of a bigger process-mining suite, not its own growth story. n8n raised $180M in October 2025 at $2.5B, then SAP took a strategic stake seven months later that doubled it to $5.2B. Revenue grew roughly 10x in a year. Three things sit under that valuation though: Open-core is a harder monetization path than it looks. The free self-hosted tier is the acquisition engine, but every dollar of paid revenue has to be earned on top of a product people can already run for free. I'll say the unpopular part: Make's UX taste is still better. It's not close, honestly. n8n's canvas is functional in the way a spreadsheet is functional. Doesn't seem to matter anymore. So the real question is whether "own your infrastructure" stays a real moat. My guess — it holds for maybe two more years before it's table stakes, not a differentiator. What do u think? #ProductManagement #ProductStrategy #Automation #AIagents

Ankush AgarwalAug 5, 2026

One idea from Marty Cagan's recent conversation with Udacity has been stuck in my head: "𝘈𝘐 𝘪𝘴𝘯'𝘵 𝘦𝘹𝘱𝘰𝘴𝘪𝘯𝘨 𝘸𝘦𝘢𝘬 𝘦𝘯𝘨𝘪𝘯𝘦𝘦𝘳𝘴. 𝘐𝘵'𝘴 𝘦𝘹𝘱𝘰𝘴𝘪𝘯𝘨 𝘸𝘦𝘢𝘬 𝘱𝘳𝘰𝘥𝘶𝘤𝘵 𝘵𝘩𝘪𝘯𝘬𝘪𝘯𝘨." The more I think about it, the more I believe we've been blaming the wrong bottleneck for years. For a long time, slow product development was almost always blamed on engineering. I've noticed that AI can help me prototype an idea in a few hours. The real time sink is still deciding whether that idea deserves to exist in the first place. That's not an engineering problem. 𝗜𝘁'𝘀 𝗮 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗱𝗲𝗰𝗶𝘀𝗶𝗼𝗻. The uncomfortable truth is that faster execution doesn't automatically create better products. It simply lets us make decisions—good or bad—at a much higher speed. That makes product judgment, customer understanding, and strategic clarity even more valuable. When everyone can build faster, the competitive advantage shifts upstream. Not to the teams that ship the most features. But to the teams that consistently make better product decisions. 𝗧𝗵𝗮𝘁'𝘀 𝘁𝗵𝗲 𝗿𝗲𝗮𝗹 𝘀𝗵𝗶𝗳𝘁 𝗔𝗜 𝗶𝘀 𝗰𝗿𝗲𝗮𝘁𝗶𝗻𝗴. 🎙️ Inspired by Marty Cagan's conversation with Udacity: https://lnkd.in/gpzjbMQr #ProductManagement #ArtificialIntelligence #ProductStrategy #ProductThinking #CustomerCentricity #Innovation #Leadership