Amid the rush to own the user interface, Glean is wagering on a quieter advantage: positioning itself as the intelligence infrastructure operating underneath it.
Seven years ago, Glean aimed to become the enterprise equivalent of Google—an AI-driven search platform built to index and retrieve information across a company’s SaaS stack, from Slack and Jira to Google Drive and Salesforce.
Today, its focus has evolved. Rather than simply developing a more capable enterprise chatbot, Glean is repositioning itself as the connective infrastructure linking AI models with enterprise systems.
Jain explained—speaking at Web Summit Qatar—that the original layer Glean built, a robust enterprise search product, required a granular understanding of employees: how they operate, how teams collaborate, and what individual preferences look like. That institutional insight, he noted, has now become the bedrock for developing high-quality AI agents.
He emphasized that while large language models (LLMs) are highly capable, they are inherently general-purpose. On their own, they lack awareness of a company’s internal structure—its people, workflows, products, and operating context. To unlock real enterprise value, he argued, the models’ reasoning and generative capabilities must be tightly integrated with proprietary organizational context.
Glean’s value proposition is that it already structures and maps that internal context, positioning itself between AI models and enterprise data.
For many customers, the Glean Assistant serves as the front door—a familiar chat interface powered by a blend of proprietary models such as ChatGPT, Gemini, and Claude, alongside open-source alternatives, all grounded in internal company data. However, Jain maintains that long-term retention depends less on the interface and more on the infrastructure beneath it.
The first differentiator is model flexibility. Rather than locking enterprises into a single LLM vendor, Glean functions as an abstraction layer, enabling organizations to switch among—or orchestrate across—multiple models as the ecosystem evolves. For that reason, Jain views OpenAI, Anthropic, and Google not as competitors but as ecosystem partners. According to him, Glean’s product improves in parallel with advances made by those model providers.
The second pillar is deep system integration. Glean connects extensively with enterprise tools such as Slack, Jira, Salesforce, and Google Drive, mapping information flows across them and enabling AI agents to execute actions directly within those environments.
The third—and arguably most critical—component is governance. Jain stressed the necessity of a permissions-aware governance and retrieval layer that not only surfaces relevant information but dynamically filters outputs based on the requester’s access rights. In large enterprises, this capability often determines whether AI remains in pilot mode or scales into full production. Simply ingesting all internal data into a model and layering controls afterward is not viable, he argued.
Mitigating hallucinations is another core requirement. Glean’s system cross-verifies outputs against source documents, provides granular citations, and ensures that responses adhere to established access controls.
The broader strategic question is whether such a middle layer can endure as platform incumbents extend further down the stack. Microsoft and Google already command much of the enterprise workflow surface area and continue to deepen their integration. If tools like Microsoft Copilot or Gemini can tap into the same internal systems under identical permission structures, does an independent intelligence layer retain its relevance?
Jain contends that enterprises are reluctant to be locked into a single model vendor or productivity ecosystem. Instead, many prefer a neutral infrastructure layer over a vertically integrated assistant embedded within one provider’s stack.
Investors appear aligned with that thesis. In June 2025, Glean secured a $150 million Series F round, pushing its valuation to $7.2 billion. Unlike frontier AI laboratories, the company does not require massive compute expenditures to sustain its growth. As Jain summarized, the business is scaling rapidly and remains financially robust.
