Enterprise AI의 '중간층'이 승리할까? Glean이 미치게 만드는 새로운 전략

Enterprise AI is turning into a chaotic battle: Microsoft is stuffing Copilot everywhere in Office, Google is pushing Gemini into every corner of Workspace, and OpenAI/Anthropic are selling directly to companies. But hold on—there’s a tiny startup called Glean that’s betting on something no one else is talking about yet.

무슨 일이 일어났나

Seven years ago, Glean set out to be the 'Google for enterprises'—an AI-powered search tool designed to index and search across a company’s SaaS tools (Slack, Jira, Google Drive, Salesforce). But today, their strategy has shifted: they’re no longer building a better chatbot. Instead, they’re aiming to become the connective tissue between large language models (LLMs) and enterprise systems.

Their CEO said at Web Summit Qatar that while LLMs are powerful, they’re 'generic'—they don’t understand your business, who your people are, or what you do. So Glean’s pitch is simple: they already map that context (your company data) and sit between the model and your systems.

First, model access: Instead of forcing companies to commit to one LLM provider, Glean acts as an abstraction layer—letting them switch or combine models. That’s why they don’t see OpenAI/Anthropic/Google as competition but partners: 'Our product gets better because we leverage their innovation.'

Second, connectors: Deep integration with systems like Slack, Jira, Salesforce to map how information flows and enable agents to act inside those tools.

Third, governance: The most important part—building a permissions-aware layer that verifies model outputs against source documents (no hallucinations) and respects access rights. This is critical for large organizations deploying AI at scale.

They raised $150M in Series F last year, doubling their valuation to $7.2B.

내가 보기엔

I’m excited about Glean’s approach—being the 'middleman' instead of a vertical assistant is smart. It gives enterprises flexibility and avoids lock-in. But I’m also skeptical: when Microsoft or Google can access internal systems with Copilot/Gemini, does this standalone layer still matter?

Think about it: if your company uses both ChatGPT and Claude for different tasks, Glean lets you switch without changing providers. That’s a huge win. But will big tech let them exist? Microsoft already controls Office, Google controls Workspace—both are pushing their own assistants deeper into the stack.

Also, governance is key here. If your company has sensitive data (like customer info), Glean’s system verifies outputs against source docs and respects permissions. That’s something even big players might struggle with if they’re too focused on speed over security.

So my guess? Glean will survive—if they keep innovating on that middle layer. But it’ll be a tight race. The real question is: who needs to win in this war? Not the big names, but the companies that can connect their data and models without being locked into one ecosystem.

Bottom line: Glean isn’t just another AI assistant—they’re building the infrastructure that makes enterprise AI actually useful. And if they succeed, every company using LLMs will thank them for keeping things flexible.