Glean: The Work AI Platform and Where WorkingAgents Fits

By James Aspinwall, co-written by Alfred Pennyworth (my trusted AI) — March 7, 2026, 07:46


Glean raised $150 million at a $7.2 billion valuation in February 2026 — their Series F, bringing total funding to approximately $770 million. DST Global led. Kleiner Perkins, Lightspeed, Sequoia, and General Catalyst participated. The company crossed $100 million in ARR with 93% adoption rates within two years of deployment. They launched an open agent platform with MCP server support, acquired Moveworks for $200 million, and shipped 50 million agentic actions in the past year.

Glean is not an enterprise search engine anymore. It is a Work AI platform — search, assistant, and autonomous agents unified by an Enterprise Graph that maps every person, project, document, and process in an organization. And they just opened the doors for external agent integration via MCP servers.

That is where WorkingAgents enters.

What Glean Does

Glean connects to all of a company’s applications and data — Slack, Teams, Salesforce, Jira, Confluence, GitHub, ServiceNow, Zendesk, Google Workspace, Microsoft 365 — over 100 connectors — and builds a real-time knowledge graph on top of it.

The Enterprise Graph

This is Glean’s core technology. The Enterprise Graph is a dynamic knowledge graph that captures:

On top of the Enterprise Graph sits a Personal Graph for each employee — their role, interests, recent activity, and contextual relevance. The combination means Glean does not just find documents. It ranks them by how relevant they are to the specific person asking.

The Product Suite

Glean Search — Semantic search across all connected applications. Not keyword matching — understanding. “Who owns the Q4 revenue forecast?” returns the person, the document, and the context, not a list of files with those words.

Glean Assistant — A conversational AI that answers questions grounded in company data. Ask it a question and it synthesizes answers from multiple sources with citations. No hallucination — every claim links back to the source document with permission enforcement.

Glean Agents — Autonomous AI agents that perform multi-step workflows. Built with the Agent Builder (no-code, branching logic, agentic looping) or via APIs using LangChain, LangGraph, NVIDIA NIM, or OpenAI Agents SDK.

Glean Protect — The security and governance layer. Centralized control over all AI activity. Data loss prevention, permission enforcement, and audit trails across every agent action.

Core Capabilities

100+ Connectors — Native integrations with enterprise applications across every category:

Category Examples
Communication Slack, Microsoft Teams, Gmail, Outlook
Productivity Google Workspace, Microsoft 365, Notion, Confluence
Engineering GitHub, GitLab, Jira, Linear
CRM/Sales Salesforce, HubSpot, Gong
Support Zendesk, ServiceNow, Intercom
Data Snowflake, BigQuery, Tableau
Storage Google Drive, OneDrive, SharePoint, Box, Dropbox

Agent Builder — No-code interface for creating custom agents with branching logic, conditional routing, and agentic looping. Agents can chain actions across multiple systems.

Agent Governance — Glean Protect enforces permissions at every step. An agent cannot access data the requesting user cannot access. Every action is logged. Administrators control which agents can access which tools.

Open Agent Platform — Glean supports bidirectional MCP communication, LangChain, LangGraph, NVIDIA NIM, and OpenAI Agents SDK. External agents can connect to Glean’s knowledge, and Glean’s agents can call external tools.

MCP Server Support

This is the direct integration surface with WorkingAgents.

Glean launched remote MCP servers in public beta — 20+ pre-loaded MCP servers (Asana, Canva, Notion, GitHub, and more) available to Glean Agents out of the box. But the platform is bidirectional:

An external agent connected to Glean’s MCP server can search company knowledge, retrieve documents, and get contextual answers — all permission-gated. A Glean Agent connected to an external MCP server can take actions in external systems.

WorkingAgents is an MCP server with 86+ tools. The connection is native.

The Numbers

Glean Value
Valuation $7.2B
Series F $150M (Feb 2026)
Total funding ~$770M
ARR $100M+
Adoption rate 93% within 2 years
Connectors 100+
Agentic actions delivered 50M+
Pre-loaded MCP servers 20+
Moveworks acquisition $200M
Key investors DST Global, Kleiner Perkins, Lightspeed, Sequoia, General Catalyst
Customers Databricks, Duolingo, Booking.com, Grammarly, Pinterest, Deutsche Telekom
Partnerships Dell, Palo Alto Networks, Snowflake, Workday
Glean:GO conference 10,000+ participants

Why This Matters for WorkingAgents

Glean knows everything about an organization. WorkingAgents does things within an organization. Glean answers “what do we know about this customer?” in seconds. WorkingAgents answers “what should we do about this customer?” with scheduled tasks, follow-ups, and escalation chains.

Glean has the knowledge graph. WorkingAgents has the operational engine. Neither replaces the other. Together, they create agents that know everything and forget nothing.

The Synergy Map

1. Enterprise Knowledge Access for WorkingAgents Agents

WorkingAgents’ chat module (ServerChat) runs agent sessions that call MCP tools to accomplish tasks. Today, those agents operate on WorkingAgents’ own data — CRM contacts, tasks, alarms, files, monitoring. They know what WorkingAgents knows.

Connected to Glean’s MCP server, those agents suddenly know what the entire organization knows:

WorkingAgents’ operational tools gain organizational intelligence. Every task, every alarm, every notification can be informed by the full knowledge of the enterprise.

2. WorkingAgents as an Action Layer for Glean Agents

Glean Agents can find information and answer questions. But what happens after the answer?

A Glean Agent discovers that a contract expires in 30 days. Now what? A Glean Agent identifies that a support ticket has been open for 72 hours with no response. Now what? A Glean Agent finds that three team members referenced the same unresolved bug in different Slack channels. Now what?

WorkingAgents provides the “now what”:

Glean Agents connected to WorkingAgents’ MCP server can discover 86+ tools for scheduling, task management, CRM operations, notifications, file operations, and monitoring. The knowledge-to-action pipeline becomes seamless.

3. MCP Server to MCP Server — Bidirectional Integration

Both platforms expose MCP servers. Both platforms consume MCP servers. This creates a bidirectional integration where:

Glean → WorkingAgents (Glean Agent calls WorkingAgents tools):

  1. Glean Agent discovers an overdue project milestone
  2. Calls task_create on WorkingAgents to create remediation tasks
  3. Calls pushover_send to notify the project manager
  4. Calls task_snooze to schedule a follow-up check in 48 hours

WorkingAgents → Glean (WorkingAgents Agent queries Glean knowledge):

  1. WorkingAgents alarm fires for a scheduled customer check-in
  2. Agent queries Glean: “Latest activity with CustomerX”
  3. Gets synthesized context from emails, CRM, support tickets
  4. Creates an informed follow-up task with full context

Combined loop (both directions in one workflow):

  1. Alarm fires in WorkingAgents → triggers agent session
  2. Agent queries Glean for context → gets organizational knowledge
  3. Agent reasons about the situation → decides on actions
  4. Agent calls WorkingAgents tools → creates tasks, sends notifications, schedules follow-ups
  5. All actions logged with audit trail in both systems

Two MCP servers. One agent. Complete knowledge-to-action loop.

4. Permission Model Alignment

Both Glean and WorkingAgents take permissions seriously — and their models are complementary.

Glean’s permission model: Mirrors source system permissions in real time. If you cannot access a file in SharePoint, you cannot find it in Glean. The Enterprise Graph enforces this at every query.

WorkingAgents’ permission model: Per-user, per-tool access control. Every MCP tool call is gated by a permission check. AES-256-CTR encrypted keys. Audit trails for every grant, revocation, and denial.

Together:

An enterprise deploying both gets permission-gated AI from knowledge retrieval to operational action. No data leakage in either direction.

5. Agent Governance and Compliance

Glean Protect provides centralized governance over all AI agent activity. WorkingAgents provides granular operational audit trails. The combination addresses enterprise compliance requirements that neither can satisfy alone:

Requirement Glean Protect WorkingAgents
Data access control Real-time permission mirroring Per-user, per-tool access control
Audit trails Agent activity logs Alarm history, task provenance, permission logs
Data isolation Enterprise Graph scoping Per-user SQLite databases
Action control Agent governance policies Tool-level permission gating
Compliance SOC 2, data residency Encrypted keys, no key serialization

For regulated industries — healthcare, finance, government — this layered governance model satisfies both “who can see what” (Glean) and “who can do what” (WorkingAgents).

6. Glean’s Customer Base as WorkingAgents Market

Glean’s customer list represents organizations already investing in AI-powered workflows:

Every Glean customer using Glean Agents for knowledge discovery needs WorkingAgents for knowledge-informed action. The pitch: “Your agents can find anything. Now they can do something about it.”

7. Moveworks Integration Expansion

Glean acquired Moveworks for $200 million — an AI copilot for IT and HR service management. Moveworks handles employee requests: password resets, software provisioning, PTO queries, IT troubleshooting. This acquisition brings operational workflow capabilities into Glean’s platform.

WorkingAgents complements this expansion:

The Partnership Path

Phase 1: MCP Server Integration

Connect WorkingAgents as an MCP server available to Glean Agents. Demonstrate the knowledge-to-action workflow: Glean Agent discovers information → calls WorkingAgents tools to create tasks, schedule alarms, send notifications. Validate permission enforcement across both systems.

Phase 2: Glean as Knowledge Source for WorkingAgents

Connect Glean as an MCP server for WorkingAgents’ agent sessions. WorkingAgents agents gain access to the full Enterprise Graph — every alarm-triggered workflow, every scheduled task, every customer follow-up can be informed by organizational knowledge.

Phase 3: Joint Agent Workflows

Build reference implementations of complete workflows: sales pipeline management (Glean for context + WorkingAgents for scheduling), incident response (Glean for runbooks + WorkingAgents for escalation chains), compliance monitoring (Glean for policy documents + WorkingAgents for scheduled audits).

Phase 4: Enterprise Reference Architecture

Co-publish an enterprise AI agent architecture: Glean for knowledge and governance, WorkingAgents for operations and orchestration. Position the combined stack for the enterprise Work AI market — agents that know everything the organization knows and reliably execute on that knowledge.

The Bottom Line

Glean built the memory of the enterprise — a knowledge graph that understands every document, every person, every relationship, and every permission boundary in an organization. WorkingAgents built the operational engine — scheduling, task management, escalation, monitoring, and crash-recoverable workflows that ensure things get done.

Glean Agents can answer any question about the organization. WorkingAgents agents can take any action within the organization. Connect them via MCP, and you get autonomous agents that are both informed and effective — agents that know what the company knows and do what the company needs.

Glean has the Enterprise Graph. WorkingAgents has the operational graph — tasks, alarms, contacts, monitoring, permissions. Layer them together and every scheduled action is informed by organizational context, every piece of knowledge can trigger operational follow-through, and every step is permission-gated and auditable.

They built the brain that remembers everything. We built the hands that get things done. Both work better together.

Sources: