It was 2019. I was at Rubrik, the data security company I co‑founded, trying to pull up an internal document from a specific project in our knowledge base. Fifteen minutes of Slack searching. Another ten in Google Drive. More time in Jira, Confluence, and the company wiki. Every tool promised to make work faster. Instead, I was burning an hour on something I knew existed.
That was the moment I looked at my screen and said out loud: “I built search systems at Google for years—why can’t I find my own company’s data?”
“Arvind Jain couldn’t find his own company’s data at Rubrik. So in 2019, before ChatGPT, before the AI boom, he built Glean, the first enterprise generative AI company.”
The irony stung. I had spent more than a decade at Google as a distinguished engineer leading teams across Search, Maps, and YouTube, and then helped build Rubrik into one of the fastest growing companies in cloud data management. Yet inside my own organization, finding information was harder than searching the entire public web. Every SaaS tool was its own silo—Slack, Google Workspace, Jira, Salesforce, Confluence, Figma, GitHub, ServiceNow. Companies had spent billions digitizing their operations, only to leave employees with disconnected fragments: a knowledge graph splintered across hundreds of apps.
I walked away from a comfortable role at a successful company—a “cozy Google job,” as some called it—to fix a problem no one realized was crippling the modern workforce. I was not chasing hype. ChatGPT did not exist yet. The phrase “generative AI” meant nothing to most people. But I understood what nobody else seemed to grasp: the most valuable AI wouldn’t train on public internet data. It would train on your company’s private data, understand your permissions, and speak your language. If you could connect the dots across an enterprise’s fractured knowledge graph, you could build something exponentially more valuable than general‑purpose search.
continue reading
English