Glean

Glean is an enterprise Work AI platform founded in 2019 by former Google Search and Rubrik engineers. It was publicly launched in 2021 and has since expanded from search into a full AI assistant and agent platform. It targets mid-market and large enterprise customers (500+ employees) and operates as a horizontal, cross-functional platform serving engineering, sales, HR, legal, and finance teams on a single deployment. Not to be confused with glean.ai, a separate accounts-payable software company.

A 100+ connector indexing layer ingests data from Google Workspace, Microsoft 365, Slack, Salesforce, Confluence, Jira, GitHub, Zendesk, and dozens more, building a proprietary Enterprise Graph that maps relationships between people, documents, projects, and processes. Search results are real-time and strictly permissions-enforced — users see only what they are authorized to see in source systems. Glean Chat provides a conversational AI assistant grounded in indexed enterprise knowledge with citations back to source documents. Glean Agents (GA as of May 2025) lets any user build, deploy, and govern multi-step AI agents in natural language to automate workflows. An Agentic Engine 2 handles deeper reasoning and orchestration. Admin-facing Insights Chat and agent observability provide usage analytics and step-by-step agent debugging. On-premises deployment is available in collaboration with Dell Technologies for air-gapped or data-sovereign environments.

Core AI capabilities include semantic vector search powered by deep learning LLMs, an Enterprise Graph for relationship-aware retrieval, a third-generation AI Assistant with personalized context, AI-generated document summaries, natural language chat with grounded citations, and Glean Agents for agentic task execution. Supports model choice — customers can connect their own LLM keys (OpenAI, Anthropic Claude via Bedrock, Gemini via Vertex AI, NVIDIA NIM, Llama) and receive FlexCredit discounts for self-hosted LLM infrastructure. REST API is well-documented at developers.glean.com with SDKs in Python, Go, and TypeScript; includes search, chat, indexing, agents, and governance APIs with an active changelog. MCP: Glean ships a fully managed Remote MCP Server built into the platform (generally available), accessible at each customer's Glean instance URL with OAuth 2.1 + PKCE and dynamic client registration. Exposes Search, Chat, Read Document, Code Search, People, and custom Glean Agents as invocable tools. Tested and documented for Cursor, VS Code, Claude Desktop, ChatGPT, and JetBrains; compatible with any MCP-compliant client via the MCP Configurator. Also publishes an open-source local MCP server on GitHub (MIT license, runnable via npx or Docker) for self-hosted or testing use cases. MCP usage is subject to FlexCredit consumption pricing. Agent interoperability additionally covers LangChain, LangGraph, NVIDIA NIM, and OpenAI Agents SDK (via MCP).

No public pricing page. All contracts require direct sales engagement. Per-user, per-month subscription model with a separate Enterprise Flex tier that introduces usage-based FlexCredits for agentic and advanced AI workloads (FlexCredits are pooled org-wide and priced by query/task complexity). Based on third-party buyer data (Vendr, GoSearch, eesel): base search reportedly starts around $45–$50/user/month; Work AI generative features reportedly add approximately $15/user/month on top. Minimum enterprise contract reportedly ~$60,000/year (~100-seat floor). Large deployments (1,000+ seats) can exceed $240,000/year in base licensing; mid-sized cloud hosting infrastructure adds an estimated $10,000+/month in additional costs for on-prem or self-hosted LLM configurations. Multi-year contracts (2–3 years) unlock better per-user rates. POC typically paid. Annual price escalators of 3–7% unless capped at signing. All figures are third-party estimates; Glean does not confirm list pricing.

100+ pre-built connectors covering Google Workspace, Microsoft 365 (SharePoint, Teams, Outlook), Slack, Salesforce, Confluence, Jira, GitHub, Zendesk, ServiceNow, Dropbox, Box, Workday, SAP, and more. Indexing API allows custom data sources to be ingested. Strategic partnerships with AWS (AWS Marketplace), Google Cloud (Vertex AI), Microsoft Azure (Bedrock/OpenAI), Databricks, Snowflake (structured data querying), Palo Alto Networks (AI security), Workday, Dell Technologies (on-prem deployment), and NVIDIA NIM. MCP Gateway allows third-party MCP servers to be proxied securely through Glean, extending its tool ecosystem to any external MCP-compatible service. Glean Agents API enables open agent-to-agent orchestration and integration with LangChain, LangGraph, and OpenAI Agents SDK.

SOC 2 Type II certified (reports available via Trust Portal under NDA). ISO/IEC 27001 certified (information security management). ISO/IEC 42001:2023 certified (AI management systems — one of the first enterprise AI platforms to hold this standard). HIPAA compliant. GDPR compliant. Zero-trust security architecture; permissions enforced at query time by mirroring source-system ACLs rather than copying data permissions into a separate store. All data encrypted at rest (AES-256, FIPS 140-2 validated) and in transit (TLS 1.2+). Single-tenant cloud deployment — each customer's data is isolated. Customer data is not used to train shared or public models. DPA available; MCP sessions covered under DPA. Bug bounty program active. Admin controls include session revocation across MCP host applications. On-prem deployment option available via Dell AI Factory partnership for air-gapped environments.

Founded 2019 in Palo Alto, CA. Founders: Arvind Jain (CEO — Google Distinguished Engineer, Rubrik co-founder), T.R. Vishwanath (Microsoft, Meta), Piyush Prahladka (Google, Uber), and Tony Gentilcore (Google). Publicly launched search product in September 2021. Total funding ~$765M across six rounds as of June 2025: Series A ($15M, 2019, Kleiner Perkins/Lightspeed/Slack Fund); Series B ($40M, 2021, General Catalyst); Series C ($100M, 2022, Sequoia, unicorn at $1B); Series D ($200M, 2024, Kleiner Perkins, $4.6B); Series E ($260M, Sept 2024, Altimeter/DST Global, $4.6B); Series F ($150M, June 2025, Wellington Management, $7.2B). ARR: crossed $100M in FY ending Jan 2025, $200M in December 2025, estimated ~$300M by May 2026 (Sacra). Notable customers include Booking.com (14,000 employees), Zillow, Ericsson, GCash, TIME, Databricks, Pure Storage. Named to CNBC Disruptor 50 (2025), Fast Company Most Innovative Companies top 10 (2025), Gartner Tech Innovator in Agentic AI. ~850 employees as of mid-2025.

Moderate. Glean is a powerful horizontal knowledge platform, but its value to a firm like a firm is narrower than its enterprise target. On the positive side: the MCP server is mature and well-documented, enabling AI-native workflows where a Claude or Cursor session can query Glean-indexed internal documents in real time; the Enterprise Graph and permission-aware search would handle internal memos, diligence notes, research docs, and meeting summaries well if connected to a firm's Google Workspace or Microsoft 365; and the compliance posture (SOC 2 Type II, ISO 27001, HIPAA, single-tenant cloud) is enterprise-grade. The friction: Glean is sized and priced for 500–14,000 employee enterprises, with a reported ~100-seat minimum contract floor (~$60K/year). A small VC or growth equity firm of 10–50 people will not meet minimum commitments without overpaying significantly. The platform's horizontal, department-spanning value proposition is also less differentiated for a firm that primarily needs deal pipeline intelligence, company/people relationship tracking, and sourcing — use cases better served by purpose-built VC tools (Affinity, Attio). Glean has no native CRM, deal tracking, or LP data model. Best fit scenario: a larger firm (50+ staff) that has already standardized on a major document/collaboration suite and wants AI-powered universal search across all internal knowledge, with MCP-enabled agent workflows.

Minimum contract size (~100 seats, ~$60K/year) makes it prohibitively expensive for small VC/GE firms. No public pricing — requires full enterprise sales cycle and paid POC. No native CRM, deal pipeline, or LP data model; it is a knowledge retrieval layer, not a system of record for investments. Pricing scales steeply with user count and agentic usage (FlexCredits add cost for AI-heavy workflows). Cloud infrastructure costs (hosting, compute) can double or triple the base licensing TCO for self-hosted or on-prem deployments. Horizontal architecture means it competes directly with Microsoft Copilot and Google Gemini for Workspace, which are embedded in tools firms already pay for. Implementation complexity is non-trivial for smaller teams without dedicated IT.