Agenten-Kompass

AGENT PROFILES

Glean Agents

Glean Technologies, Inc. · Agent builders · Agent workspaces

Build, manage and use agents grounded in connected company knowledge for multi-step tasks.

Typical tasks

Availability in Europe
01

What it does

5 features

Glean Agents · Enterprise agent platform

Product and access

Glean Agents is part of the Glean Work AI platform with an agent builder, orchestration and governance.

Core features

5 features from provider sources · reviewed on 12 October 2026. Availability and limits are listed below.

  • Build agents

    The agent builder supports knowledge-grounded agents and multi-step workflows.

    Source 2

  • Use company knowledge

    Agents draw context from connected company sources.

    Source 1

  • Handle events

    Events can trigger agents and route tasks between them.

    Source 1

  • Respect permissions

    The platform provides permissions-aware governance and oversight.

    Source 1

  • Observe usage and errors

    Analytics show usage, error rates and feedback, among other measures.

    Source 1

Usable AI models

Fixed choice across providers

Admins decide which models from the Glean catalogue are available; in Agent Builder these are the ones enabled for agents. This works through Glean's key or your own keys, which are then tied to your cloud. The checked pages do not mention custom endpoints or local models.

Model families named

  • OpenAI (GPT)
  • Google (Gemini)
  • Anthropic (Claude)
  • Amazon (Nova)
  • DeepSeek
  • Z.ai (GLM)
  • Moonshot (Kimi)
  • Nvidia (Nemotron)
Model choice
Automatic or your choice
Your own API key
Yes
Router or custom endpoint
Not stated
Local models
Not stated
Depends on
Admin approval

The rating covers the generally available main access. Documented · Sources accessed on

Glean: Supported LLMs · Glean: Model choice

02

Does it fit my task?

2 practical tasks · 3 requirements and limits

A good fit if …

  • your organisation wants agents to use knowledge from many connected apps
  • admins should control which write actions run in the background and who publishes agents
  • you need scheduled agents with run history and audit logs

Less suitable if …

  • you want to start without a contract or demo; pricing is on request only
  • you are an individual without admin structure, since the controls assume roles

Editorial assessment based on provider documentation. Not a product test.

Tasks Research Organize office work Automate

Practical tasks

Editorial examples based on documented features. No product tests performed.

01

Create an internal briefing

Your input
A topic and approved company sources.
Possible result
A briefing grounded in internal sources.
What to check
Check source freshness and recipient permissions.
02

Coordinate processing

Your input
A trigger and defined steps in connected systems.
Possible result
A coordinated workflow with traceable handoffs.
What to check
Check failures, approvals and logs.
EDITORIAL EXAMPLE

A briefing grounded in internal sources.

Your possible prompt

Compare three providers for [task]. Use their official documentation. Show differences, sources, and open questions in a table.

Adjust the task, data and permissions.

Requirements and limits

  • Requirement

    Enterprise access

    Entry is through a demo and sales; data connections require setup.

    Source 3

  • Requirement

    Prepare data and permissions

    Agent context depends on connected sources and their access permissions.

    Source 1

  • Not yet confirmed

    Contract and allowances

    License scope, allowances and prices must be clarified with the provider.

    Source 3
03

Can I use it?

Availability still unclear

Europe & regional availability

Official product and regional sources reviewed. The stated open points still require confirmation from the provider.

Reviewed scope: Product access, feature limits and regional evidence
Reviewed on:

  • Availability still unclear

    Europe: specific availability unconfirmed

    Glean documents agents and regional partner sales. Some features depend on deployment: memory is available for GCP with Glean Universal Key; AWS and Azure require additional readiness approval. A complete European country matrix remains unconfirmed.

    Region, reason and source

    Region covered: EU/EEA, United Kingdom and Switzerland

    Documented reason: The reviewed sources do not provide a complete list of supported countries for this.

    Official source · www.glean.com · Official source · docs.glean.com · Official source · www.glean.com

Europe and regulation explained

Accounts & permissions

Based on provider documentation. Permissions depend on the account, integration and organization policies. Review the permissions actually requested before use.

Configuration covered: Glean Agents in an enterprise workspace, interactive and scheduled runs

Sources reviewed:

Show all 3 points

  • Documented with conditions

    Account used

    By default an agent runs under the identity of the user who invokes it. Alternatively an admin can set up an agent identity with a registered service credential, so the agent runs with the same fixed permissions for every invoker.

    Evidence · Evidence

  • Documented

    Admin approval

    Admins control which write tools may run in the background without confirmation; a setting for default members governs who can share and publish agents. Moderators review agents before publication. An admin must register service credentials in the admin console before builders can attach them to agents.

    Evidence · Evidence

  • Documented with conditions

    Data access & team sharing

    Access is enforced at execution time against the identity of the person executing the agent, not its creator. With an agent identity, the admin sets the credential's scope, which is the maximum access. Moderators can limit who may subscribe to an agent.

    Evidence · Evidence
04

Memory and conversation continuity

3 of 6 criteria documented

Saved and daily learned memories complement short-lived workflow context.

User personalization and separate agent-step memory

Sources accessed:

Documentation by criterion

StoragePartly documented

Storage and scope

Glean stores user-specific saved and learned memories; specific database technology is not named.

SourcesMemory personalization · Memory administration
SelectionDocumented

Selection and timing

Explicit remember requests save directly; learned memory updates daily from activities and conversations when enabled.

SourcesMemory personalization · Memory administration
Edit & deleteDocumented

Inspect, edit and delete

Users can inspect, edit and delete memories or request forgetting; deleted content is excluded from future updates.

SourcesMemory personalization
Access boundariesDocumented

Agent and channel boundaries

Personalization is private per user; agent access depends on configuration. Workflow-step state has separate boundaries.

SourcesMemory personalization · Agent memory concepts
RetrievalPartly documented

Retrieval

Answers can show used memory sources; step memory passes preceding outputs by default. Long-term ranking remains unspecified.

SourcesMemory personalization · Agent memory concepts
New conversationsPartly documented

Session reset and compaction

Saved memory remains until deletion; workflow context is separately bounded. A common conversation-compaction rule is undocumented.

SourcesMemory administration · Agent memory concepts
Memory and conversation continuity: fundamentals
05

Cost and effort

Prices as of 12 October 2026

Prices and included usage

Pricing checked

Price on request; no public amount in the reviewed sources.

  • The reviewed product and demo pages provide no specific plan price.

Pricing details

Setup

Contact sales, set up enterprise access and connect data sources.

Operation & oversight

Based on public provider documentation. Assessments are editorial inferences from the stated properties. Actual time requirements and reliability have not been measured.

Deployment covered: Cloud platform; scheduled background runs and interactive use considered

Sources reviewed:

Editorial assessment

Operating it needs admin roles for background permissions and service credentials, moderators for publishing, and someone who watches runs and audit logs. Users must reactivate deactivated subscriptions individually.

Show all 6 points

  • Documented with conditions

    Files and working state

    Each subscription keeps its own saved inputs, which every scheduled run uses. The page does not say whether conversation context carries over between runs.

    Evidence

  • Documented with conditions

    Sign-ins

    Certain configuration changes deactivate subscriptions, and each subscriber must manually reactivate the agent. A service credential does not depend on anyone's session; an admin can replace it with new credentials or revoke it, after which the agent loses access immediately.

    Evidence · Evidence · Evidence

  • Documented

    Starting further runs

    Agents start manually or on a schedule with start date, time, time zone and repeat frequency. Each user can have up to 10 active background agents; the exact start time can vary when many agents are scheduled together.

    Evidence

  • Documented with conditions

    Human intervention

    Scheduled runs have no per-run confirmation. Write tools that need confirmation are blocked from background runs. Moderators review inputs, tools and triggers before publishing.

    Evidence

  • Documented

    Detecting errors

    Runs appear under Chats with the Agents or Automations filter, and details can be opened. Admin audit logs and event logs exist, and audit logs can be streamed to SIEM tools. Subscribers receive emails about agent changes and an in-app banner when saved inputs need updating.

    Evidence · Evidence

  • Open in reviewed sources

    Resuming after failures

    The reviewed page does not describe automatic retries. Deleted or blank inputs are passed with their original value as context; scheduled runs have a maximum end-to-end duration. Paused subscriptions are reactivated manually.

    Evidence
06

Evidence and open questions

3 sources · 2 open points

Open points

The reviewed sources contain no reliable information on these points. A gap says nothing about the product.

Europe: Availability still unclear

Requirements and limits: not yet confirmed

Sources and evidence

Official product documentation reviewed. Workflows and assessments are editorial; hands-on product tests are outside the scope of this directory.

Editorial team: Michael Meierhoff · sources reviewed on . Provider information is not independent evidence of performance.

Read profile as Markdown

How we assess the sources

Provider

Provider of the Glean platform for company knowledge and AI agents.

Operator
Glean Technologies, Inc.
Organization
Company
Provider sourcesSource 1 Reviewed