Agent Orchestration Platform

7/29/2026

This post defines and categorizes agent orchestration platforms, positioning Glean as an enterprise solution. It details the technical considerations for such platforms, including workflow control, non-linear logic, state, retries, recovery, multi-agent coordination, security, permissions, approvals, observability, auditability, and integrations. It provides a comparative table of various platforms based on these criteria.

7/28/2026

This post expands on the 'Agent Orchestration Platform' by detailing the technical components that constitute Glean's 'system of context' beyond just indexing. It introduces specialized indexing strategies (semantic, lexical, structured) for different data types, the role of the Enterprise Graph in mapping relationships, the function of Memory for persistent learning, the flexibility of Data Connectors for varied data sources, and the necessity of Tools for AI action. It highlights how these elements synergize to provide AI with a deeper understanding of the enterprise, improving accuracy, latency, and token efficiency.

7/23/2026

This post details how Glean's tool stack was adapted for voice interactions, exposing and addressing limitations in retrieval quality, tool contract detail, and execution reliability. Key improvements include a cleaner, stricter tool stack with better filtering, ranking, and canonical execution paths. The system now prioritizes dynamic skill and tool discovery, using a filter-rank-trim funnel. Lexical weighting was found to be effective for disambiguating repetitive tool names. For voice, tool contracts are now more explicit and inline, including canonical names, server identity, and argument schemas. Error handling has been enhanced with structured retry signals. The execution path for action-oriented work is now more deterministic, with explicit authorization state and confirmed downstream results. The post also highlights how voice exposed existing infrastructure issues like caching and pooling, which were then addressed to improve real-time system reliability.

7/22/2026

This post details how to optimize token efficiency in agentic systems by treating it as a workflow-design problem rather than a prompt-editing problem. It identifies five key areas of token waste: retrieval noise, bloated working context, unstructured memory, over-orchestration, and unbounded loops. The post proposes solutions including measuring token usage by workflow stage, reducing context before it reaches the model through selective retrieval and context engineering, replacing raw transcripts with structured state, routing tasks by complexity, designing loops with clear stopping rules, and reusing proven plans and stable context. It also outlines key metrics for evaluating token efficiency and common mistakes teams make.

7/14/2026

This post introduces agent identity as a core component of Glean's independent agents, allowing agents to act through their own service account credentials. This decouples agent invocation from user access, provides agents with their own presence across surfaces (Slack, Jira, GitHub), ensures agents are trusted with exactly what the job needs via scoped credentials, and enables scheduled work that doesn't silently expire due to user token expiration. Actions are properly attributed to the agent's own account in the audit trail.

7/13/2026

This post details how the Enterprise Knowledge Graph serves as a foundational layer for various ROI-driving applications, including semantic enterprise search, AI/LLM grounding (GraphRAG), fraud & risk detection, supply chain visibility, content & knowledge management, regulatory compliance, and drug discovery. It highlights how the graph integrates with existing systems, provides context through relationships, enables interoperability without migration, and speeds up data access. The post also outlines stakeholder roles and a pragmatic adoption roadmap for knowledge graphs.

7/8/2026

This post introduces Glean's 100% programmatic tool-calling agent harness, which uses code execution within a sandbox for orchestration, filtering, looping, and branching. This approach reduces token usage by 24% compared to previous methods by avoiding multiple LLM round-trips and serializing/deserializing tool inputs/outputs. Key technical details include using the shell as the only directly exposed tool, exposing other capabilities via a tools SDK, employing a sandbox filesystem for tool outputs, and using tool truncation with previews to manage context. The harness also leverages search and progressive disclosure to dynamically load tools and skills, appending search results to the conversation to preserve the prefix cache.

7/6/2026

This post details the core technologies powering AI assistants, including Natural Language Processing (NLP) and Natural Language Understanding (NLU), Machine Learning (ML), Data Integration, and Speech Recognition. It explains the workflow of an AI assistant from capturing user input to understanding language and intent, reasoning and learning, executing actions across systems, and delivering results. Key features such as learning, personalization, proactive assistance, automation, voice and text support, integrations, scalability, data analysis, and security are highlighted. The post also categorizes different types of AI assistants (personal, task-oriented, conversational, voice, business, healthcare, decision-making, specialized) and outlines their benefits and use cases across industries.

6/29/2026

This post details Glean's approach to token-efficient deployment of its enterprise AI platform, focusing on optimizing retrieval, context selection, and orchestration to minimize waste and maximize outcome per token. It introduces evaluation criteria for token efficiency, including cost per task, latency per task, output quality and groundedness, and scalability. Key technical aspects discussed are passage-level evidence retrieval, context management (deduplication, permission respect, source preference), intelligent model routing across different model tiers, and multi-step orchestration with scoped working memory and intermediate result compression. The post also highlights the importance of observability and governance for measuring and managing token efficiency, and identifies red flags in vendor evaluations related to model-centric discussions, oversized context, lack of context narrowing explanation, uniform model path usage, and poor visibility into spend.

6/29/2026

Introduces independent agents as a new class of AI coworkers built on Glean's context layer. These agents enable autonomous execution across enterprise applications (Jira, Slack, Teams) with a focus on security and governance. Key features include: identity with context (provisioned access, distinct profiles), memory (learning from company data and interactions), proactivity (acting without explicit prompts), and accountability (traceable actions, emergency stop). The OnCall Assistant is presented as a specific implementation for engineering on-call escalations.

6/24/2026

This post details the integration of Glean's AI coworker agents into Slack, highlighting how they leverage the existing unified index of company knowledge and permission enforcement to provide context-aware and proactive assistance. It emphasizes the technical challenges and solutions related to ensuring agents understand what they are allowed to know and on whose behalf they are speaking, by enforcing real permissions at index time. The post also discusses how the agents utilize the organization's knowledge graph to answer complex questions spanning multiple systems and how they can proactively chime in and close the loop by executing actions like opening pull requests or updating tickets, all while maintaining a full audit trail and respecting existing permission structures.

6/18/2026

This post details how Glean's AI platform achieves token efficiency by focusing on retrieval quality, context selection, and orchestration design, rather than solely on prompt engineering. It introduces a four-layer framework: retrieving better and less, passing only relevant evidence through intelligent chunking and passage selection, structuring prompts for control, and orchestrating intelligently across steps by summarizing intermediate outputs and avoiding redundant context loading. The post emphasizes measuring token efficiency through metrics like tokens per successful task, latency per task, cost per task, retrieval precision, answer quality, and context utilization rate, highlighting common mistakes like over-reliance on large context windows or optimizing for cost alone.

6/5/2026

This post differentiates AI assistants from AI agents, emphasizing the autonomous, goal-driven, and multi-step execution capabilities of AI agents. It details how agents leverage tool orchestration and persistent memory to perform complex tasks, contrasting this with the reactive, prompt-response nature of assistants. This contributes to the understanding and application of AI agents within the broader platform.

6/2/2026

This post details how to build AI chatbots for enterprise teams using Glean, focusing on the foundational requirements of trusted retrieval, permission awareness, multi-tool context, actions, escalation paths, and pre-rollout evaluation. It highlights Glean's context system, APIs (Search API, Chat API), and Glean Agents as tools for building these capabilities. The post outlines a 5-step process: starting with a specific team workflow, ensuring the knowledge layer is correct, choosing the right build path and action scope, testing before scaling, and rolling out by team.

5/31/2026

This post details how to build a generative AI stack for software engineers, emphasizing a layered approach with coding assistants, a context platform, observability AI, and agent platforms. It highlights Glean's role as a context platform that connects code, tickets, incidents, and docs, providing permission-aware context to AI agents. The post also discusses the importance of trust, security, and fitting tools into existing workflows.

5/18/2026

This post introduces the Health Agents platform, a system of proactive, dedicated agents built on top of Glean Agents for monitoring system health across customer deployments. It details the architecture, including specialized component health agents and a custom orchestration/aggregation layer, and discusses design tradeoffs such as specialization and the separation of monitoring and debugging agents. The post also covers mechanisms for building user trust, such as high-quality data, explainability, and feedback loops, and highlights the impact of Health Agents on GTM and Engineering workflows, including proactive customer health monitoring and cross-deployment visibility.

5/17/2026

This post details how Glean's AI coworker capabilities extend beyond information retrieval to executing end-to-end workflows and managing tedious tasks. It highlights the ability to create Jira tickets with attachments and schedule meetings from scattered feedback, triage enterprise communications (Gmail, Slack, Asana), and generate personalized first drafts for sales collateral. The post emphasizes the role of the Enterprise Graph and Personal Graph in providing context for accurate and secure execution of these tasks, including data-backed decision making through querying across structured and unstructured data, invoking agent sandboxes for Python execution, and ensuring security controls like per-session isolation and permission-aware retrieval. It also touches upon model agnosticism and the ability to create various artifacts like images, slides, documents, and emails.

5/13/2026

This post describes the emerging AI tooling landscape for software engineers, highlighting a two-layer model: an enterprise context and knowledge layer, and coding and interaction surfaces. It positions Glean as a platform for the context layer, enabling AI-driven workflows for onboarding, coding, and incident response by unifying data from various sources like code repositories, tickets, and documentation. The post details how this context layer integrates with coding assistants and other tools to provide grounded, permission-aware AI outputs.

5/12/2026

This post details a benchmark comparing Glean's context layer against off-the-shelf Model Context Protocol (MCP) tools, demonstrating Glean's superior preference rate (~2.5x) and token efficiency (~30% less tokens consumed). It highlights the challenges of federated search, such as increased tool calls, latency, and potential for inaccurate results, contrasting it with Glean's centralized indexing and knowledge graph approach. The evaluation focused on utility, correctness, completeness, and tool fidelity across various enterprise tasks, showing Glean's advantage, especially in complex scenarios.

5/11/2026

This post introduces the Agent Development Lifecycle (ADLC) as a structured approach to building and managing enterprise AI agents, mirroring the software development lifecycle. It details the stages of the ADLC: Opportunity, Design, Performance, Context, Develop, Launch, and Monitor & Improve. New Glean capabilities are introduced to support the ADLC, including generally available auto-mode agents for autonomous planning and execution, debug and trace views for visibility into agent runs, sub-agents for modularity, an agent sandbox with a file system and code interpreter for secure and complex tasks, an agent library for discoverability and governance, agent access policies for enforcing guardrails, and agent insights for monitoring performance and collecting feedback.

5/5/2026

This post details how Glean's AI agents are being applied to improve sales workflows by connecting AI to fragmented enterprise data sources (CRM, call recordings, email, etc.). It highlights specific agent use cases like 'account snapshot' for sales call prep, agents for CRM hygiene and post-call follow-up, and agents for deal execution and expansion planning. The post emphasizes the technical challenge of unifying context across disparate systems to reduce friction and improve sales productivity, aligning with the platform's goal of making AI useful within real workflows.

5/4/2026

This post details how Glean's agent orchestration platform provides the context layer for an engineering AI stack. It highlights the importance of context across various enterprise systems (code, tickets, incidents, docs, chat) for AI tools to be effective. Glean's capabilities include building an enterprise context graph, providing hybrid lexical + semantic search, enforcing permission-aware access, acting as an MCP server for IDE integrations, and enabling agents to convert tickets/threads into PRs. The post also discusses evaluation criteria for AI tools, focusing on context awareness, trust earned through grounding and source linking, data security and governance, and workflow integration.

5/4/2026

This post introduces Glean Skills, a new capability within the agent orchestration platform. Skills allow for the encapsulation of reusable workflows, including instructions, logic, example outputs, and tool usage, to provide structured, consistent, and context-aware AI assistance for enterprise tasks. Skills leverage Glean's existing understanding of enterprise context and can be imported from external ecosystems. The post highlights how Skills accelerate enterprise AI transformation by moving from generic AI to reusable, on-demand expertise.

5/4/2026

This post introduces several new capabilities for Glean's AI coworker functionality. It details 'Skills' as personalized packages of reusable execution logic for automating repetitive tasks. It also describes the ability to delegate and execute multiple workstreams simultaneously using 'write tools' that can edit data across various enterprise applications, with previews before execution. New connectors are mentioned for Apollo, Dropbox, Enterpret, Gainsight, and Granola. A new feature called 'Adaptive Reasoning' is introduced, which automatically selects the appropriate AI model based on query complexity. Real-time voice capabilities for direct conversation with Glean are also highlighted, along with inline revision review in Canvas to show additions and deletions. The 'Library' is presented as a shared repository for AI-generated outputs.

4/28/2026

This post details significant advancements in Glean's agent orchestration platform, focusing on context management and harness engineering. Key contributions include the implementation of Programmatic Tool Calling (PTC) within sandboxes to move workflow logic into code, enabling more robust and efficient execution of complex tasks. The introduction of sub-agents allows for the decomposition of agent loops into isolated execution contexts, improving scalability and reliability for large-scale analysis. Compaction strategies have been developed to preserve essential conversation state while offloading intermediate outputs to file systems, mitigating context window limitations. Furthermore, a search-first discovery mechanism for skills has been implemented, decoupling skill discovery from schema hydration and improving the agent's ability to select relevant capabilities efficiently.

4/27/2026

Introduced Waldo, an agentic search model that acts as a specialized front-end to frontier LLMs. Waldo is trained to break down queries, select tools, and gather evidence before handing off to a frontier model, significantly reducing latency (~50%) and token usage (~25%). The training process involves Direct Preference Optimization (DPO) on tool-use patterns and reinforcement learning with outcome-based rewards. Waldo's execution also informs adaptive reasoning, dynamically determining the required reasoning level for a task.

4/15/2026

This post introduces a two-layer model for AI tooling in engineering, emphasizing the need for a foundational context layer alongside coding surfaces. It details how enterprise context platforms, AIOps/observability tools, and general-purpose AI platforms work together. The post highlights the evolution of AI from experiments to infrastructure, the importance of blending tools rather than vendors, and the critical role of security, governance, and explainability in building a scalable and trustworthy AI stack for engineering teams.

4/9/2026

This post introduces MCP Apps as an extension to the Agent Orchestration Platform, enabling the embedding of rich, partner-authored UIs directly within Glean Assistant. This allows users to execute work across integrated applications (e.g., Asana, Hex, Gamma, Box) without leaving the Glean interface, moving beyond simple callable actions to full application integration. Examples include creating Asana tasks from conversations and analyzing Hex dashboards within Assistant.

4/7/2026

This post introduces new capabilities within the agent orchestration platform: slide generation for presentations (e.g., PowerPoint, Google Slides) and interactive page generation (e.g., dashboards, infographics, visualizations) using HTML. These features leverage the existing permissions-aware knowledge graph and can incorporate company-approved templates for branding. Image generation and editing are integrated for visual content within presentations. The interactive pages are dynamic and grounded in live data from connected applications.

4/1/2026

This post details how Glean integrates generative AI (for content creation, summarization, and hypothesis generation) with predictive AI (for validation, forecasting, and risk assessment) within its enterprise AI stack. It explains the synergy between these two AI types, highlighting how predictive AI grounds generative outputs in fact, accelerating time-to-insight and reducing workload. The post discusses practical applications, technical considerations like unified data infrastructure and hybrid evaluation metrics, challenges such as bias and governance, and best practices for adoption. It positions Glean as a platform that enables these hybrid AI interactions across knowledge assets.

3/31/2026

This post details the seven core components of an AI agent architecture: Goal Definition, Perception and Input Processing, Memory Systems, Reasoning and Planning, Tool Execution and Action, Orchestration and Coordination, and Feedback and Observability. It explains how each component contributes to building scalable, compliant, and trustworthy AI systems, highlighting technical aspects like input sanitization, hybrid memory models, various reasoning strategies (rule-based, chain-of-thought, tree-of-thought, Monte Carlo), action sandboxing, supervisor-worker hierarchies, and observability for debugging and compliance. Glean's platform is mentioned as an example for several of these components, particularly in unifying perception, providing context-aware access to enterprise knowledge, and ensuring secure enterprise integrations.

3/31/2026

This post details Glean's approach to trace learning for self-improving AI agents. It introduces the concept of traces as records of task execution and explains how trace learning enables agents to learn from successes and failures to improve future performance. The post outlines a two-component system: offline learning, which mines strategies from historical traces using a teacher-student agent comparison, and online application, which retrieves and applies learned strategies at runtime. It emphasizes the importance of accuracy in enterprise environments, detailing a stricter validation process involving cross-response assertion checks and conflict verification. The post also highlights the design choice to store intentionally narrow tool strategies that generalize across related tasks, focusing on workflow-level patterns rather than individual tool calls. Finally, it explains how learnings are split into deployment and user levels for secure capture of individual preferences, with actions replayed in a sandbox for safety.

3/25/2026

This post clarifies the role of the Model Context Protocol (MCP) in enterprise AI, distinguishing its value at scale from local workflows. It highlights how remote MCP servers act as a central layer for enterprise AI, enabling centralized per-user auth and secrets management, org-wide telemetry and observability, standardized prompts/skills and docs across surfaces, and a single integration surface for many hosts. This extends the agent orchestration platform by providing a robust, scalable mechanism for AI agents to interact with enterprise tools and data in a governed and observable manner.

3/19/2026

This post introduces the concept of combining no-code automation tools with context-aware AI agents, positioning Glean's AI agents as a complementary layer that addresses the limitations of traditional no-code solutions. It highlights how AI agents can leverage deep enterprise context (knowledge in docs, tickets, historical activity, relationships) to reason, adapt, and act, thereby enhancing complex, knowledge-driven workflows. The post contrasts event-driven, rules-based no-code tools with goal-driven, context-aware AI agents, emphasizing the latter's ability to understand intent, pull relevant context, and decide which tools to use. It positions Glean's agents as filling the gap where no-code tools hit a wall due to rigid logic, shallow reasoning, and lack of memory or learning.

3/18/2026

Introduces LLM choice as a feature within Glean Assistant, allowing users to select specific Large Language Models (LLMs) for different tasks. This involves integrating with various LLM providers and enabling administrators to curate and govern model usage. The post details how this flexibility improves task performance and user confidence, while also providing administrative controls for oversight and compliance.

3/12/2026

This post details how Glean's agent orchestration platform is applied to the energy and utilities sector. It highlights the use of AI agents to address industry-specific challenges like demand supercycles, aging infrastructure, and workforce shifts by connecting fragmented data systems. Specific applications include GenAI for planning & forecasting, regulatory & compliance, engineering & field service, customer service & support, and training & onboarding. The post also emphasizes the platform's security features, including a zero-trust architecture, real-time permissions mirroring, and data governance, to meet the stringent requirements of regulated industries.

3/9/2026

This post defines and elaborates on the concept of 'context engineering' as a critical discipline for building reliable and high-performing AI systems and agents. It details the components of context (content, structural, task, activity layers) and the inputs to an LLM's context window (system instructions, user input, short-term memory, long-term memory, RAG, tool calls/outputs, structured output formats). The post emphasizes how effective context engineering leads to higher accuracy, faster outputs, more aligned decision-making, lower costs, and safer AI by providing models with curated, structured, and relevant information, differentiating it from prompt engineering.

3/9/2026

This post details the integration of the Model Context Protocol (MCP) into Glean's agent orchestration platform. It introduces MCP as a standard for connecting AI models to enterprise systems, enabling agents to perform actions beyond answering questions. Key technical contributions include: the implementation of a remote MCP server, the expansion of the MCP tools ecosystem within Glean Assistant and Agents, the development of tool search for efficient action discovery from a large set of MCP tools, the use of enterprise memory to capture and surface successful action sequences, and the definition of reusable skills as executable markdown specifications for complex workflows. The post also elaborates on the security enhancements for MCP, including centralized governance, least-privilege execution, human-in-the-loop verification, and runtime scanning for prompt injection and malicious code.

2/17/2026

This post introduces over 85 new actions for Glean's agents, expanding their integration with external applications like Salesforce, Jira, GitHub, Confluence, and Google Calendar. These actions enable agents to perform complex tasks such as creating/updating records, searching with application-specific query languages (SOQL, JQL), reviewing code, and generating documentation. The post details how these actions work in practice with a support engineer workflow example and highlights administrative controls for enabling and configuring action packs.

2/17/2026

This post introduces the concept of agent sandboxes as a solution to the limitations of LLM context windows for long-running agent tasks. It details how sandboxes provide agents with a virtual environment (file system, command line, code runtime) to act as short-term memory, enabling them to process larger datasets and perform complex analyses. The post explains how sandboxes integrate with Glean's Enterprise Graph and leverage a Python runtime for data processing. It also discusses lessons learned from previous memory management techniques, such as data classification and selector tools, and how these inform the sandbox design. Finally, it outlines the security measures implemented for sandboxes, including isolation and access controls.

2/16/2026

This post details significant enhancements to Glean Assistant's agent orchestration capabilities. Key new features include: agent sandboxes for processing large datasets and complex tasks, a suite of new native actions for interacting with enterprise tools (Salesforce, Jira, Confluence, GitHub, Google Calendar), content generation (documents, images, slide decks, code), and improved personalization through user-editable personal graphs. It also introduces real-time voice support and new task management agent templates (daily action items, plan my day, weekly work snippets). Canvas interface has been updated with adaptive UIs for Slack and email.

2/11/2026

This post details an evaluation comparing Glean's enterprise context layer against ChatGPT's company knowledge and Claude's enterprise search. It highlights Glean's superior performance (1.9x preference over ChatGPT, 1.6x over Claude) due to its specialized indexing, Enterprise Graph, and enterprise memory, which effectively address context rot and improve retrieval accuracy. The post contrasts Glean's approach with the 'deep' (sequential, in-depth) strategy of Claude and the 'wide' (parallel, broad) strategy of ChatGPT, explaining how Glean's architecture provides more accurate and relevant context for AI agents.

2/8/2026

This post introduces Glean's support for the Agent Skills standard, a reusable package of instructions, scripts, and tools that teach AI how to tackle specific tasks. It details how Glean Assistant uses these skills for tasks like account planning, and provides tips for skill crafting, including the importance of negative examples and few-shot examples in skill instructions. The post also discusses the symbiotic relationship between skills and agents, and outlines a future where Glean automatically builds skills based on user work patterns.

2/8/2026

This post details the evolution of Glean's agent architecture, emphasizing a horizontal, open strategy. It highlights the importance of context engineering, including data connectors, indexes, and knowledge/context graphs, as a foundational layer for reliable enterprise agents. The post also discusses the shift towards a multi-model, multi-provider future, the role of orchestration in automating workflows, and the critical need for unified security across agent deployments. It introduces the concept of context graphs for understanding enterprise workflows and the feedback loop between context and orchestration for continuous improvement.

2/4/2026

This post details Glean's approach to building context graphs, which are essential for AI agents to understand and automate work processes. It explains that context graphs model the 'how' of work by transforming actions into first-class entities with temporal traces and causal/correlational edges. The construction involves deep connectors and observability to capture change events, building a unified knowledge graph to infer entities and relationships, and creating a personal graph to understand individual tasks. Aggregating personal graphs and normalizing them into abstracted traces forms the context graph, which represents a probabilistic view of work flows. The post also highlights the hybrid storage model used for event data and the importance of closing the loop with agent execution, where agent traces are incorporated into the context graph for continuous learning and refinement. The context graph is presented as a joint model of human and agent behavior, owned by both the data and orchestration layers.

1/21/2026

This post details Nextdoor's implementation of Glean's agent orchestration platform to drive Work AI adoption. It highlights the creation of a 'learning loop' involving office hours, Slack channels, bulletins, and in-product guidance. Specific technical contributions include connecting core employee systems (e.g., Google Drive, Slack, Confluence, Coda) to Glean for seamless information access and action, implementing permission-aware context, and showcasing diverse agent use cases such as an open enrollment guide (Finn), instant client intelligence for sales, Trust & Safety analysis, on-demand technical references for engineers (improving productivity by 2-3x), and automated address enrichment for RevOps. The post also mentions the use of SSO, Slack integration, a Chrome extension, and a no-code agent builder to reduce friction, and the establishment of an agent directory and micro-workshops for curation and reuse.

1/14/2026

This post introduces seven new marketing-focused AI agents that leverage the Glean agent orchestration platform. These agents demonstrate the platform's capability to ingest and process various data sources (sales calls, articles, internal documents, event briefs) to generate specific marketing outputs such as SEO keyword research, article evaluations, customer testimonials, customer reference summaries, event messaging briefs, event descriptions, and LinkedIn post drafts. The post highlights how these agents can be used individually or in conjunction to support end-to-end marketing workflows, emphasizing the platform's role in transforming scattered context into usable outputs.

12/10/2025

This post introduces eight new quickstart sales agents that leverage the agent orchestration platform for sales productivity. These agents automate tasks such as account snapshot generation, prospect outreach email drafting, deal strategy synthesis, competitive brief creation, intelligent reminders, sales call coaching, deal loss insights analysis, and account handoff documentation. The agents are designed to pull data from various systems (CRM, Slack, email, call transcripts, news) and provide actionable, contextualized output for sales teams.

12/9/2025

This post details the evolution of context engineering within Glean's AI platform, shifting the burden from individual engineers to a unified, continuously learned, and consistently delivered platform capability. It introduces specialized indexes for different data types (calendar, documents, structured data), emphasizes the importance of connectors and data modeling for heterogeneous enterprise data, and highlights the role of knowledge graphs for multi-hop reasoning. The post also describes advancements in tool search and enterprise memory for optimizing agent actions, including storing and learning from agent run traces to improve tool selection, parameterization, and action sequencing.

11/20/2025

This post details key metrics for evaluating enterprise search tools in 2025, focusing on accuracy, speed, user experience, and pricing. It introduces specific benchmarks for tool calling accuracy (90%+), context retention (90%+), and response times (<1.5-2.5 seconds). It also discusses update frequency, interface usability, reporting quality, and compares Glean with other leading search tools based on these metrics. The post emphasizes aligning search tool capabilities with business needs and use cases, particularly for AI-powered search platforms.

11/16/2025

This post details the integration of Glean with Microsoft Agent 365, extending Glean's agentic capabilities into Microsoft Word, Outlook, and Teams. It highlights how this integration leverages the Microsoft 365 Model Context Protocol (MCP) server to enable cross-surface actions and personalized experiences. The post also emphasizes the security aspects, including Glean running in the customer's Azure cloud, enforcement of Microsoft sensitivity labels, and protection against AI-specific threats.

11/9/2025

This post details the application of Glean Agents within the Agent Orchestration Platform to empower legal teams. It highlights how these agents leverage Glean's core capabilities (indexing, Enterprise Graph, Data Connectors) to provide AI-driven assistance for tasks such as contract review, redlining, and regulatory tracking. The post emphasizes the importance of grounding AI in verified data, enforcing real-time permissions, and providing auditable outputs. It also introduces specific customer examples demonstrating measurable improvements in efficiency and cost savings through the use of these agents.

11/6/2025

This post details how Glean's Work AI platform enables AI coworkers to assist manufacturing teams by providing shared context across siloed data systems (ERP, PLM, MES, QMS, etc.), facilitating faster decision-making, improved coordination, and proactive risk management. It highlights capabilities like enterprise-grade RAG over structured and unstructured data, native and custom connectors for unified context, security mirroring source permissions, an Assistant + Agents plane for search-to-execution, and incremental adoption patterns. The post also discusses integration strategies, build vs. buy considerations, and ROI through pilot programs and automation of workflows like incident response and handoffs.

11/4/2025

Introduces a new surface for Glean's AI Assistant: a native desktop application. This app provides full context across all open windows, unlike browser extensions or mobile apps. Key new features include: 1. Instant keyboard shortcuts (Command-Shift-J) to summon the Assistant without leaving the current application. 2. An 'always-on-top' quick chat window for side-by-side interaction and iteration. 3. Screenshot-to-chat functionality, allowing users to share visual context (documents, UI states) for more relevant answers. The post also reiterates Glean's commitment to privacy and control, stating the app respects enterprise permissions and does not index local files.

11/3/2025

Introduces a unified agent builder that merges graph-based and conversational approaches, allowing for natural language creation and step-by-step building. Enhances the plan & execute step with 'Fast' and 'Thinking' modes to optimize for speed or depth, leveraging an upgraded Agentic Engine 2 with increased reasoning power and adaptability. Implements more granular agent permissions for better control over agent sharing and access. Introduces agent routing from Assistant to automatically direct user queries to specialized agents based on configured rules.