Arvind Jain left a decade at Google and Rubrik to build Glean after finding no enterprise search product to buy for Rubrik's own knowledge-sharing problem.
Jain skipped market sizing and customer validation, sending thousands of low-response LinkedIn messages and deliberately not charging for two years, rejecting the Eric Ries Lean Startup playbook.
Glean moved from vitamin to painkiller once users averaged more than ten searches a day and saw ChatGPT-style tools, and Pablo Srugo says revenue grows three to four times a year.
Glean's founder ignored lean startup rules and built a $2bn company
Glean shows a founder succeeding by backing a problem he had lived with conviction, instead of following the Lean Startup playbook. Jain saw knowledge scattered across ever more SaaS tools and found no product to fix it. He built for two years before charging, and the arrival of ChatGPT-style tools made the need urgent.
Takeaways
Glean grew from a problem Jain lived at Rubrik. Rubrik passed 1,000 people in four years and tripled engineering, yet productivity fell. Surveys showed staff could not find information or know whom to ask. Jain found no enterprise search product to buy, so he built one.
SaaS sprawl makes enterprise search a growing need. Companies now run hundreds or thousands of applications, which fragments their knowledge. Jain says company knowledge doubles every year and spreads across 30% more systems, so the problem keeps getting worse.
Jain skipped validation, market sizing and lean iteration. He did no TAM analysis, avoided business-minded people who would talk him out of it, and compared search's unprovable return to email's. Because he believed in the problem, he saw no need to pivot.
Glean built free for two years before charging. Jain was the sole business development rep for two years, sending thousands of LinkedIn messages with low response. Glean charged only once users averaged over ten searches a day, and the team grew to about 20.
The product was hard to build and became urgent with AI. Glean had to connect to hundreds of systems with clunky APIs, normalize data, remove stale information and handle semantics. It used BERT-family models from day one. Jain says ChatGPT-style tools turned the vitamin into a painkiller.
Growth and funding followed once Glean started selling. Pablo says Glean raised $15m at the start, then a $40m Series B in 2021, then $200m at a valuation above $2bn. Its first sales year brought $3m, with revenue growing three to four times annually.
Conviction and persistence matter more than market analysis. Jain tells founders to solve a problem they believe in and ignore competitors, market size and doubters. He says the real failure rate is nearer 99 in 100, because founders drop ideas before launching.
Jain and Pablo on the Lean Startup method
Pablo says building for two years without charging breaks Eric Ries's Lean Startup approach of fast iteration and minimum viable products. He says the method is so ingrained that almost everyone runs it blindly, and it may not fit every case.
Jain says Glean had a fundamental belief in the problem and did not need to pivot. He ties the difference to founder goals, since first-time founders often want to succeed as entrepreneurs while he wanted to solve a problem for people.
Tap a name to go there · tap a line to see why they are linked