Arvind Jain did not start Glean because he wanted to run a search company. He started it because he could not buy the product he needed.

He had been one of the early search engineers at Google, then left to co-found Rubrik. Rubrik grew quickly, and the growth produced a problem he had not planned for. Writing on Glean's blog in 2025, he described hitting a productivity wall after the company passed 2,000 employees in under four years, because nobody could find the information or the experts they needed. Rubrik's own employee survey put it plainly: not being able to find a document or the right colleague was the biggest productivity complaint in the company. Engineers spent too much time outside code. Account managers could not find the research they needed to close deals. New hires took too long to get up to speed.

His first move was to buy a fix rather than build one. He looked for an enterprise search product and, in his words, could not find anything that met their needs. When he asked around, other companies reported the same thing, and studies he read put the average employee at roughly a day a week spent looking for information. That was the moment the problem stopped being a Rubrik problem.

Glean was founded in March 2019 by Jain with T.R. Vishwanath, Piyush Prahladka and Tony Gentilcore. That month is the anchor for every number below. Jain has written that they started Glean in 2019, and in a post on 11 March 2026 he noted that the day coincided with Glean's seventh birthday, which puts the founding in March 2019.

Then they disappeared for more than two years. When Glean finally went public on 15 September 2021, Jain's LinkedIn post opened with "After over two years building in stealth." The product was not a demo at that point. His launch day blog post talked about customers whose employees were saving two hours per week, and Glean's own company timeline says the product was already in use at more than 40 companies on the day it came out of stealth. The two years of silence were spent selling quietly to a small set of companies and building against what those companies actually had installed.

The technical bet was the interesting part. Jain has written that earlier attempts at enterprise search fell short because they relied on federated search, which queries whatever search APIs each application happens to expose. That approach misses content and knows nothing about how people, documents and activity relate to each other. Glean took the Google approach instead: crawl continuously, index everything, normalize it, and rank it. On top of that they added a knowledge graph of relationships inside the company, a lexical algorithm rebuilt for enterprise data, and a per company language model built on BERT that learns each company's internal vocabulary. Jain says search quality typically improves about 20 percent in a customer's first six months for that reason.

There was one more piece that made the timing work. In the 2021 post he pointed out that the same SaaS sprawl causing the problem had also standardized it. Because workplace apps now expose permissions and ranking signals through APIs, Glean could connect to a company's stack in under two hours instead of the months an older integration project would have taken.

Then the market moved toward them. ChatGPT arrived in 2022, every CIO wanted large language models inside their company, and the models had no company context and made things up. Retrieval augmented generation fixed that by grounding answers in search results, and Glean had spent three years building exactly the retrieval layer that approach needs. The company raised a $100 million series C at a $1 billion valuation, which Jain wrote was just over half a year after coming out of stealth, then over $200 million at a $2.2 billion valuation in February 2024, then over $260 million at $4.6 billion later that year.

On 5 February 2025 Jain posted the number this story is anchored on. Glean had surpassed $100 million in ARR in its last fiscal year, and he described the pace as "in just a little over three years." Measured from the March 2019 founding, that is 2,158 days. Measured the way Jain measures it, from the point Glean started selling, it is a little over three years. Both numbers matter: the revenue clock ran fast, the company clock ran slow, and the two years nobody saw are the reason the fast part was possible.

The engagement numbers he published alongside it explain the retention. Glean users average five queries a day, about the same as consumer web search, and the weekday DAU to MAU ratio sits near 40 percent against the 10 to 20 percent typical of enterprise software. Named customers at the time included Deutsche Telekom, Booking.com, Reddit, Zillow and LinkedIn. In a founder bylined post published on the Glean blog in December 2025, Jain reported that ARR had passed $200 million, nine months after crossing $100 million.

For anyone building something smaller, the useful part is not the funding. It is the sequence. Jain solved a problem he had personally measured inside his own company, confirmed with other companies that it was not just his, checked whether he could buy the answer before deciding to build it, and then took two years to build the unglamorous layer underneath. When the market shifted, he did not have to rebuild anything. He already had the part everyone else suddenly needed.