The enterprise AI land grab is on. Glean is building the layer beneath the interface.

The conflict for endeavor AI is heating up. Microsoft is bundling Copilot into Office. Google is pushing Gemini into Workspace. OpenAI and Anthropic are selling straight to enterprises. Every SaaS vendor present ships an AI assistant. 

In the scramble for the interface, Glean is betting connected thing little visible: becoming the quality furniture beneath it. 

Seven years ago, Glean acceptable retired to beryllium the Google for endeavor — an AI-powered hunt instrumentality designed to scale and hunt crossed a company’s SaaS instrumentality library, from Slack to Jira, Google Drive to Salesforce. Today, the company’s strategy has shifted from gathering a amended endeavor chatbot to becoming the connective insubstantial betwixt models and endeavor systems.

“The furniture we built initially – a bully hunt merchandise – required america to profoundly recognize radical and however they enactment and what their preferences are,” Jain told TechCrunch connected past week’s episode of Equity, which we recorded astatine Web Summit Qatar. “All of that is present becoming foundational successful presumption of gathering precocious prime agents.”

He says that portion ample connection models are powerful, they’re besides generic. 

“The AI models themselves don’t truly recognize thing astir your business,” Jain said. “They don’t cognize who the antithetic radical are, they don’t cognize what benignant of enactment you do, what benignant of products you build. So you person to link the reasoning and generative powerfulness of the models with the discourse wrong your company.”

Glean’s transportation is that it already maps that discourse and tin beryllium betwixt the exemplary and the endeavor data. 

The Glean Assistant is often the introduction constituent for customers — a acquainted chat interface powered by a premix of starring proprietary (ie, ChatGPT, Gemini, Claude) and open-source models, grounded successful the company’s interior data. But what keeps customers, Jain argues, is everything underneath it. 

First is exemplary access. Rather than forcing companies to perpetrate to a azygous LLM provider, Glean acts arsenic the abstraction layer, allowing enterprises to power betwixt oregon harvester models arsenic capabilities evolve. That’s wherefore Jain says helium doesn’t spot OpenAI, Anthropic, oregon Google arsenic competition, but alternatively arsenic partners. 

“Our merchandise gets amended due to the fact that we’re capable to leverage the innovation that they are making successful the market,” Jain said. 

Second are the connectors. Glean integrates profoundly with systems similar Slack, Jira, Salesforce, and Google Drive to representation however accusation flows crossed them and alteration agents to enactment wrong those tools. 

And third, and possibly astir important, is governance. 

“You request to physique a permissions-aware governance furniture and retrieval furniture that is capable to bring the close information, but knowing who’s asking that question truthful that it filters the accusation based connected their entree rights,” Jain said. 

In ample organizations, that furniture tin beryllium the quality betwixt piloting AI solutions and deploying them astatine scale. Enterprises can’t simply load each their interior information into a exemplary and make a wrapper to benignant retired the solutions later, says Jain. 

Also captious is ensuring the models don’t hallucinate. Jain says its strategy verifies exemplary outputs against root documents, generates line-by-line citations, and ensures that responses respect existing entree rights. 

The question is whether that mediate furniture survives arsenic level giants propulsion deeper into the stack. Microsoft and Google already power overmuch of the endeavor workflow aboveground area, and they’re bare for more. If Copilot oregon Gemini tin entree the aforesaid interior systems with the aforesaid permissions, does a standalone quality furniture inactive matter?

Jain argues enterprises don’t privation to beryllium locked into a azygous exemplary oregon productivity suite and would alternatively opt for a neutral infrastructure furniture alternatively than a vertically integrated assistant.

Investors person bought into that thesis. Glean raised a $150 cardinal Series F successful June 2025, astir doubling its valuation to $7.2 billion. Unlike the frontier AI labs, Glean doesn’t request monolithic compute budgets.

“We person a precise healthy, fast-growing business,” Jain said.