At Glean, we’ve spent the past few weeks preparing for our annual user conference, Glean:GO. During this time, I’ve been reflecting on how far AI has come since last year’s conference. Yet, even with the incredible strides we’ve made as a platform for enterprise AI, and the dramatic changes we’ve made as an industry, there are important questions to answer before we can claim AI has reached its potential:
- How do we control surging token costs and make AI value predictable?
- How do we manage AI sprawl and balance rapid innovation with the governance and security businesses demand?
- How do we stop botsitting and make AI that acts without having to be prompted every time?
In my experience, robust, comprehensive, well-governed context is the answer to each of these questions. Access to the right information, where and when it’s needed, significantly improves AI efficiency and costs. When AI systems can understand the totality of your business, governance and security become first-class citizens because you have full visibility when driving rapid innovation. Finally, with a platform built on context, you have the confidence that AI can accurately tackle proactive work without human prompting and intervention.
At Glean, we’ve been working for years on delivering the context, efficiency, and governance needed for the next generation of AI. We’re building on that foundation as the leading Enterprise AI platform and shaping how AI reaches its next phase of potential.
Context maximizes enterprise value alongside cost efficiency
Many AI front doors rely on shallow context. They leverage MCP which pulls fragmented information across data sources and lean on frontier models to overcompensate for the lack of information, resulting in models using more reasoning loops to fill knowledge gaps. In both these situations, outputs suffer and token costs rise. Glean drives efficiency by providing the right context upfront, minimizing repeated searches and reasoning loops.
We’re announcing Glean Intelligence, a new layer of our platform that works alongside Glean Enterprise Context to maximize enterprise value with model choice, auto routing, and AI usage controls.
- Model hub has always provided access to leading models in the market without lock-in. With 40+ models, we have recently added NVIDIA Nemotron 3 Ultra and GLM 5.2 as open models to our portfolio, and will be adding Kimi K3 and DeepSeek v4 soon.
- Auto routing automatically matches AI requests to the right model in real time, maximizing the intelligence needed for a task while consuming the minimum tokens.
- Glean’s AI usage controls enable customers to manage model access and control usage across their organization.
In our most recent benchmark that compared Glean and Claude Cowork across 180+ tasks for sales, engineering, marketing, finance, customer support, and human resources, Glean saved 81% on token costs and was preferred 78% of the time. This translates into a Glean average cost per query of $0.58, with Claude Cowork coming in at $2.98. One of the reasons that Glean won is that it's not locked into a single model or provider. It routes across models to match each query to the right one, instead of paying frontier-model prices for every task by default.
.jpg)
In a market defined by rapid innovation and relentless model competition, enterprises benefit most when they have the flexibility to quickly and easily implement the latest models. When adopting a new model, within every organization, there is a tradeoff between intelligence and cost that you are trying to optimize. The place where the tradeoff is perfectly optimized is what we call the Pareto frontier. You can't just chase the most capable model for every task. You need the right model at the right cost for the right job.
Auto routing takes into account the tasks and complexity to determine the optimal user experience to balance price and performance. In addition, Glean layers in usage data to understand how close a user is to their limits. If a user is reaching their capacity, Glean will auto route them to a more token-efficient model. Auto routing comes with built-in control, so admins can choose and prioritize cost efficiency or frontier performance when routing models to their organization’s work.
.jpg)
While Glean helps you optimize your intelligence and efficiency with auto routing, you still need to understand, monitor, and control access and usage across your organization. Glean’s AI usage controls enable you to set monthly usage limits across your organization, users, departments, and agents. You can diagnose what product surfaces drive department usage, and understand how individual users are using different models, agents, and products. Users can then request additional capacity when needed, and you can ask Glean to dive deeper into what’s driving particular usage while forecasting future spend to avoid surprises.
Leverage the power of Glean’s context and platform, no matter where work gets done
As AI spreads across more systems and teams within an organization, the landscape gets more complex, including more vendors, models, and front doors into your business. The result is that context and model usage become fragmented, causing gaps in governance, limited cost control visibility, and inconsistent work every time a new AI surface appears. At Glean, we believe the answer is to fight that sprawl by building the right foundation with quality context, consistent security, model choice, and openness.
Glean provides a single, unified AI gateway that gives enterprises one control plane for AI across the front doors they use. With Glean’s AI Gateway, organizations can manage model access, govern tool execution, and security policies in one central place, instead of rebuilding those controls app by app.
Glean connects its permissions-aware enterprise context to the tools people already use. Last year, we announced Glean’s remote MCP server; since then, 78% of customers have adopted Glean MCP and APIs—a clear signal that enterprise context is becoming a high-value asset. And context is more than data: it increasingly includes personal memory and organizational skills, too. With Glean, memory follows users where they work, learns preferences across surfaces, and stays centrally governed.
Security teams increasingly need to distinguish normal agentic activity from genuine threat patterns. Glean Protect now empowers these teams with context-aware threat detection, which uses Glean’s Enterprise Graph to connect actions—including those involving sensitive data—to the user behind them, enabling detection of data exfiltration, data harvesting, and mass destructive write operations. Glean’s Enterprise Graph understands users and their work to be able to detect at higher confidence whether the behavior is risky or business-appropriate.
No matter where your teams work together, our aim is to give you what you need for AI to take on work across your enterprise, including high-quality context, security across every front door, and the openness and model choice to do it cost-effectively.
Context powers proactive AI: the next generation of AI work
Today, we’re announcing the next generation of AI grounded in Glean’s platform of context, intelligence, and protection – proactive AI. Built for individual and team productivity, proactive AI continuously observes business signals, understands patterns, and acts before asking. It redefines productivity by reducing reactive work and embedding powerful capabilities directly into how and where organizations operate.
We’re also announcing Glean Tau, Glean’s new desktop app that combines local files, apps, and code with Glean’s enterprise context. Built for code execution and coworking, Glean Tau makes coding accessible to all your employees so they can delegate complex, long-running tasks across authorized tools. Glean’s AI Gateway provides access to the models used, governs which tools and MCP servers are available, manages how requests get routed, and applies the appropriate security and governance policies. And because Glean Tau is configured right out-of-the box, your employees get all the models, MCP servers, skills, and memory with one click.
At Glean, engineering teams use Glean Tau for coding and technical troubleshooting, while finance teams update large spreadsheets and build financial models in Excel using natural language. Other teams use it to turn multimodal inputs, including audio transcripts stored locally, into concise business analyses and presentations. By combining local files with Glean Enterprise Context and tools that can take action securely through Glean’s AI Gateway, Glean Tau becomes a single place where work starts and gets done across the business.
Proactive AI moves work forward where communication and personal productivity typically break down. Across managing tasks, calls, and emails, Glean’s proactive AI plays an active role in every touchpoint of how work moves forward, prioritizing work, advising you, and executing on your behalf.
- Glean’s new task management capabilities leverage the personal graph to connect conversations, decisions, and commitments, while identifying and executing specific next steps. It drafts follow-up messages and moves projects forward by completing a first draft of the deliverable before you ask.
- Email triage prioritizes your inbox by identifying what deserves attention and taking a first pass at the response or next action
- Meeting coach delivers personalized preparation and in-call guidance grounded in the people, projects, past discussions, and context around the conversation.
Interactive dashboards, powered by enterprise context, deliver deep, up-to-date insights capable of guiding organization-wide decision making. Interactive dashboards use Glean’s agent harness to analyze structured and unstructured data, like tickets, conversations, and documents, providing a near real-time view of critical information across your organization, not just what's in your data warehouse. Interactive dashboards refresh on a schedule or when a dashboard is viewed. By reflecting the dynamic nature of all the data, touchpoints, and people within your organization, it ensures content is fresh enough to proactively guide on-the-spot decision making.
Glean’s new independent agents proactively guide and coordinate how entire teams work together. With their own identity that includes their own profile, scoped service credentials, permissions, and audit trail, independent agents become valuable team members who can work autonomously across departments. They combine real autonomy with strong governance and security, so they can act across important workflows like on-call management, voice of the customer, and sales without losing control. They learn from company context to build standard operating procedures, embed into tools like Slack, Jira, GitHub, and more, and maintain their own identity and memory over time, allowing them to get up to speed quickly and proactively act as a coworker. Because they carry context forward, they improve with every interaction and feel less like a one-off prompt and more like a persistent, trusted teammate.
In addition to collaborating with independent agents, employees and Glean can proactively work together to create artifacts with team chat, bringing context and a multiplayer experience together in a shared chat thread. When a user begins a conversation, teammates can join the same chat session with Glean and work together using the same context. You can send messages, discuss the work, and build on the context already surfaced. Sharing is controlled with the ability to review and revoke access at any time, and Glean responds using the active participant’s permissions.
Glean Transform uses Glean's enterprise context to map work in real time, so you can see, by job function and job family, where time is being spent, what automation opportunities exist across your organization, and how AI has been changing work there. Glean proactively finds opportunities for automation and defines workflows across every department, across all employees, and aggregates them at a job family level, to ensure that individual users' activities are kept private. Using this workflow understanding, Glean can suggest skills and agents to take on critical parts of business processes, driving better results. Glean Transform also provides your organization with a comprehensive look at the business results driven by these changes, so you can have proof of AI's impact.
Context is the foundation of enterprise AI
From my conversations with customers, it is clear that AI breaks down without a strong foundation of enterprise context – costs rise, sprawl takes hold, and AI stays static. As we look to the future, context will continue to sit at the center of everything. Context lowers costs while improving performance by surfacing the right information at the right time. Glean brings that context everywhere people work, securely governed through a single platform. Context powers proactive AI, observing business signals, recognizing patterns, and acting on high-impact work before it’s asked. We’re excited about what this future will make possible and I can’t wait to build it alongside our customers.
Interested in learning more about how Glean’s taking the next step and bringing out the full potential of AI with comprehensive, well-governed context? Sign up for Glean:GO and get a front seat.
Generally available: Glean Intelligence, AI usage controls, interactive dashboards, memory via MCP
Beta: Independent agents, AI Gateway, auto routing, team chat, interactive dashboards live and scheduled refresh
Coming soon: Glean Tau, Glean Transform, usage requests, spend forecasting, self-serve usage analysis and reporting, context-aware threat detection, task management, meeting coach, email triage, skills via MCP, Kimi K3, and DeepSeek V4
The decision to purchase our services should be made based on features that are currently generally available.
