Clarifying questions

Glean Assistant can ask short, contextual follow-up questions when your request is ambiguous or missing key details. These clarifying questions appear only when needed, not for every prompt. They ask for your judgment only when Assistant needs it to confidently infer the right path from available context.

How clarifying questions work

1

Assistant detects ambiguity

When you submit a task-execution request, Assistant evaluates whether key details such as audience, scope, format, or emphasis are missing or ambiguous.

2

Interactive clarifying questions appear

Assistant displays one to three questions, each with two to four clickable options and a free-text field labeled Something else for custom answers. A Skip button is always available.

3

Assistant proceeds with your input

After you answer or skip, Assistant incorporates your selections and begins the task. If you skip, Assistant proceeds with reasonable defaults. Your responses and any skipped questions are reflected in the output.

Example

For example, you ask Assistant to "Create a customer-ready presentation about our AI assistant capabilities."

Before starting work, Assistant displays two questions:

1. Which narrative should I anchor on? with options such as Assistant-first, Platform-first, and Use-case-first.

   

2. For audience and depth, which style should I optimize for? with options such as Executive first call, Business + technical buyers, and Technical deep dive.

   

When clarifying questions appear, Assistant shows a small set of suggested responses so you can answer quickly. You can choose one of the suggested options or type your own response in the free-form text box.

Why clarifying questions help

Clarifying questions improve the quality of Assistant's output in several ways:

• Better results on the first try. When Assistant generates a document, presentation, or analysis with the wrong framing, each revision can take minutes. Clarifying questions resolve ambiguity upfront so the first output is closer to what you need.
• Less wasted work. A quick chat answer that misses the mark takes seconds to fix with a follow-up. A presentation or report that takes minutes to generate with wrong assumptions wastes significantly more time and effort. Clarifying questions prevent that.
• A collaborative experience. Instead of guessing your intent and hoping for the best, Assistant works with you to align on scope, audience, and format before it starts.
• Accessible for everyone. Clickable options reduce cognitive load so you don't need to write a detailed, expert-level prompt. You can guide Assistant's work with a few clicks.
• Context-aware questions. Questions adapt to the type of task. A slide-creation request might ask about audience and key takeaway, while a data analysis request might ask about time range and metrics.

When you may see clarifying questions

Clarifying questions appear for task-execution requests where Assistant needs your judgment before producing a substantial artifact or taking action. This includes work such as drafting a document, creating a presentation, taking action on an enterprise data source, or running a multi-step analysis.

You won't see clarifying questions for information-seeking requests like searching for a document, asking a factual question, or getting a quick answer. For these queries, Assistant proceeds immediately with the best answer it can find.