glean glean

Create an Enterprise Knowledge Search Agent

Build and deploy an AI‑powered agent that instantly searches your company's documents, apps, and data sources to answer employee questions with accurate, contextual replies.

Takes
Text Document Url
Produces
Text Document
What you'll produce
Input
How many support tickets did we close last week?
Output
We closed 124 support tickets last week, as shown in the support dashboard.

The Workflow

9 steps · click a step number to mark it done

Open glean and have it ready before you start

Identify the query class you want the agent to handle, e.g., employee FAQ, ticket metrics, or product documentation.

Connect Glean to your internal data sources by adding document repositories, APIs, and app integrations in the Data Sources settings.

Create a new Agent in Glean and select the appropriate LLM model for reasoning and generation.

Define the search tool configuration: map query intents to specific data sources and set any required advanced search operators.

Design the workflow steps—search, retrieve, format, and present—using Glean's visual workflow builder.

Add context selection rules so the agent includes relevant metadata (team, department, time range) in each prompt.

Test the agent with sample queries, review the generated answers, and adjust search relevance or formatting as needed.

Enable access controls so the agent respects user permissions when pulling data from protected sources.

Publish the agent to your organization’s chat or portal and monitor usage analytics to iterate and improve.

Did this workflow help?
1
🔍 Looking for AI tools? Try searching!