A New Chapter in Work AI: From Prototype to Production
San Francisco, CA – May 20, 2025 — At its annual Glean:GO event, Glean unveiled one of the most anticipated advancements in enterprise AI: the general availability of Glean Agents, a powerful, open, and enterprise-grade platform to operationalize AI agents across the workplace.
After months of strong anticipation, Glean has delivered not just a product, but a blueprint for the future of intelligent work. With today’s launch, the company moves decisively beyond agent experimentation — bringing automation, insight, and orchestration into the hands of every employee, securely and at scale.
“AI agents should be in the hands of every employee and driving outcomes across the entire enterprise,” said Arvind Jain, Founder & CEO of Glean. “Glean Agents turns that vision into a reality with an open, interoperable platform that integrates across systems, supports leading LLMs, and makes it easy to build and scale agents that actually move the business forward.”
A Platform Built for the Enterprise, From the Ground Up
Unlike many agent solutions on the market today — which are limited to narrow copilots or single-use tools — Glean Agents offers a horizontal, full-stack framework that supports the entire lifecycle of AI agents:
- Build with a natural language agent builder or developer APIs
- Deploy across any business function
- Orchestrate complex workflows through agent-to-agent collaboration
- Govern securely with enterprise-grade observability, permissions, and compliance
Key Capabilities Unveiled at Glean:GO
1. Agent Creation for All
At the core of the platform is Glean’s natural language agent builder — a simple, no-code interface that allows employees to create agents by describing their goals in plain English.
For developers and technical teams, Glean provides a robust Agents API for fine-grained control, programmatic orchestration, and integration into external systems.
2. Model Choice with No Lock-In
Glean Agents is model-agnostic, giving enterprises the freedom to select from 15+ leading large language models (LLMs) via a universal model key — spanning OpenAI, Claude, Mistral, Gemini, and models from Amazon Bedrock, Google Vertex AI, and Azure OpenAI.
Even more powerful: users can assign different models to different steps within a single agent, optimizing for accuracy, latency, or cost.
3. Agent Interoperability at Scale
A major highlight from the event was the introduction of Glean’s hosted Model Context Protocol (MCP) server, enabling agents to securely collaborate across systems and platforms — a foundational step toward scalable, interoperable multi-agent environments.
This architecture also supports agent-to-agent coordination, where one agent can call another or delegate tasks — laying the groundwork for dynamic, self-assembling workflows across departments.
Prebuilt Agents, Ready to Deliver Impact
Glean isn’t just delivering a platform — it’s launching with momentum. Included in today’s release:
- 30+ Quickstart agents across Sales, IT, HR, and Engineering
- Personal productivity agents for tasks like meeting prep, action summaries, and performance reviews
- A deep research agent that blends internal and external data for contextual, citation-backed analysis
- Structured query agents for Snowflake Cortex, enabling natural language analytics grounded in live enterprise data
This combination gives organizations the ability to accelerate time-to-value, reduce build friction, and demonstrate ROI early in their agent journey.
Security, Governance, and Observability — by Default
Recognizing the risks and compliance challenges that come with autonomous agents, Glean has engineered its platform with security at the core:
- Prompt injection and jailbreak protection baked into the agent execution layer
- Sensitive content detection across more than 100 integrated enterprise systems
- Granular access controls for creation, viewing, and editing agents
- Step-by-step observability and evaluation tools, enabling builders to trace agent behavior, optimize logic, and ensure reliability
“Success depends not just on innovation, but on control, interoperability, and enterprise alignment,” said Andrew Gens, Research Analyst for AI Software at IDC. “Platforms like Glean’s that address these needs will be essential in moving pilot projects to production-grade AI adoption.”
Strategic Collaborations Expand Ecosystem Value
Today’s announcement also highlighted key partnerships that further extend the reach and value of Glean Agents:
- Workday: Glean and Workday will enable agent-to-agent collaboration across both platforms, streamlining HR and Finance workflows.
- Dell Technologies: Glean will offer its full stack — including enterprise search and agents — for on-premises deployment, addressing compliance needs in regulated industries.
- Snowflake: A new integration with Snowflake Cortex allows agents and users to run real-time natural language queries over structured data, closing the loop between insight and action.
- Palo Alto Networks: A joint effort to secure AI agent interactions across hybrid data environments, giving IT leaders enhanced visibility and control.
Availability
Glean Agents is available starting today, including:
- Natural language builder and agent library
- Quickstart agents and personal agents
- Model hub and universal model key
- Per-step model and temperature selection
- Snowflake Cortex integration
- Agent API (beta)
- MCP Server (beta)
- Agent observability and eval tools (beta)
- On-prem architecture with Dell (available now)
Looking Ahead
Today’s release isn’t just about new technology — it’s a signal that AI agents are ready to enter the core of enterprise operations.
With a secure, scalable foundation and a commitment to openness, Glean is helping organizations move past the “AI pilot phase” and into real, transformative deployments.
For enterprises ready to make the leap from experimentation to execution, Glean Agents offers a path forward — and a platform built to last.
To learn more, visit: glean.com/agents
Stay tuned as I continue exploring this evolution on The Ravit Show — breaking down agent architectures, enterprise adoption frameworks, and what’s next in the agentic era.
