Enhancing Your Experience with Glean

The following instructions detail how to optimize your experience using the Glean Assistant. We'll walk through the process of utilizing Glean's enhanced features to ensure you can access the most accurate and comprehensive information from your company's internal resources, web content, and AI capabilities.

Step 1

We're thrilled to introduce the upgraded Glean Assistant experience. The first feature to explore is the All Knowledge mode.

Step 2

All Knowledge mode integrates your company's internal knowledge, web content, and AI's built-in knowledge to provide the best response to your query. Based on your query, Glean will determine whether it should search the web or focus on internal resources. You can customize this by toggling off "Search the Web" to focus solely on internal resources or by toggling off "Use Company Resources" to prioritize web searches.

Step 3

If both options are toggled off, Glean will utilize the LLM model, specifically GPT-5, to address your query.

Step 4

Another significant update is the differentiation between Fast and Thinking modes.

Step 5

Leveraging GPT-5, Glean offers remarkable reasoning and thinking capabilities to integrate into daily workflows. Fast mode is ideal for straightforward, quick queries and routine information retrieval. In contrast, Thinking mode engages the LLM for more intricate reasoning and complex problem-solving, employing the appropriate tools for precise answers.

Step 6

Thinking mode caters to more complex workflows or queries within Glean. Let’s demonstrate with a straightforward example using Fast mode.

Step 7

For a simple inquiry like, "What is our PTO policy?" Fast mode suffices as it does not require extensive reasoning. It retrieves information from the available data corpus, offering comprehensive guidelines regarding the PTO policy.

Step 8

The response includes detailed instructions on requesting time off, exceeding the initial query's requirements.

Step 9

To discover information such as "What are the public holidays in Singapore?" Glean can be queried directly.

Step 10

Glean will check internal resources and perform a web search if necessary, demonstrating its ability to execute searches, such as retrieving data from the Ministry of Manpower website.

Step 11

It successfully compiles an official list from the Ministry of Manpower, detailing Singapore's public holidays in 2025.

Step 12

Glean effectively discerns the appropriate search path based on your query.

Step 13

Now, let's explore when to use Thinking mode.

Step 14

Thinking mode is designed for complex queries necessitating deeper reasoning and understanding to achieve specific outcomes.

Step 15

For example, when preparing troubleshooting documentation based on a Zendesk ticket, Thinking mode is advantageous.

Step 16

You can direct Glean to extract resolution steps from the ticket, generate an RCA, and compose customer communication. This complex query is best served by Thinking mode.

Step 17

Upon issuing this query, Glean undertakes multiple tasks, engaging in thoughtful reasoning to address the task comprehensively.

Step 18

Glean employs document reading, crafts troubleshooting steps, summarizes incident timelines, analyzes snippeting structures, and delivers a thorough response. Introducing reasoning does create slight latency, but it's worthwhile for the detailed output.

Step 19

Glean reviews the Zendesk ticket and drafts complete documentation, including an RCA and customer communication, providing a timeline, impact summary, and resolution steps, and even initiates a troubleshooting runbook.

Step 20

What typically takes hours across multiple platforms like Zendesk, Jira, and Slack, Glean accomplishes in under a minute. This is the power of the latest updates to your Glean Assistant experience. If you have questions, feel free to connect on the Gleanverse.