What a $7B founder learned building Glean from scratch

Executive overview

Enterprise employees lose hours daily to information they can't find. Arvind Jain saw this at his own company, Rubrik, and built Glean to fix it — an AI layer that connects to all company data so anyone can ask questions and get answers instantly.

The key bet: don't build foundational models. Use what already exists and focus engineering effort where no one else is paying attention.

Build for a problem you live every day, then hold conviction long enough for the world to catch up.

The founding insight

• Rubrik's per-person productivity metrics deteriorated as headcount grew
• Employees cited information-finding as their biggest blocker
• The problem was universal — every person Jain spoke to agreed immediately
• Leaders weren't used to buying enterprise search, but the problem still existed

Battling enterprise search fatigue

• Past enterprise search products had failed universally — buyers were skeptical by default
• Jain spent early months cold-outreaching on LinkedIn; most went unanswered
• Sending cold outreach as a founder with no product built resilience and fortitude
• High quality was non-negotiable: the benchmark users expected was Google-level search

Validating before charging

• Offered the product free for the first two years to 20 design partners
• Usage signals were unambiguous: security teams tried to shut it down; employees revolted
• Sequoia came to Glean — Jain didn't pitch them; word of mouth did it
• Customers had already internalised that charging would come; the GA transition was smooth

Engineering strategy: don't reinvent the wheel

• Glean deliberately does not train its own foundation models
• Leverages GPT, Claude, Gemini — enterprise customers prefer this approach
• Focus goes entirely to the hard, unsolved layer: bringing enterprise context to AI tasks
• As a pure-play enterprise AI company, no legacy products compete for attention

Advice for aspiring founders

• Validate the problem with enough people, then stop seeking permission
• Investors and early hires will be lukewarm — that is normal
• The reason you started thinking about the problem still exists; stay true to it
• Don't become the person who kills your own idea