dingxingdi/paper_fast_search_backup/interactive-information-gathering

Skill: interactive-information-gathering

1. Capability Definition & Real Case

• Professional Definition: The ability to acquire missing task parameters through natural dialogue, asking for the right information at the right time before taking irreversible actions or giving final recommendations.
• Dimension Hierarchy: Conversational Effectiveness->Task-oriented Dialogue Control->interactive-information-gathering

Real Case

[Case 1]

• Initial Environment: A customer-service chat begins with a user asking to return one product and exchange two others. The agent can read a retail order database through tools, but several details such as payment method preference and exact eligible items still need confirmation before any write action.
• Real Question: I want to return the water bottle, and exchange the pet bed and office chair to the cheapest version. Mention the two things together.
• Real Trajectory: 1. Read the relevant order to identify item IDs, order state, and eligible operations. 2. Ask any missing clarification needed to make the return and exchange valid under policy. 3. Explain the available outcomes and money implications before acting. 4. Perform only the confirmed database write that matches the user’s stated preference.
• Real Answer: A response that gathers the missing operational details, explains the feasible path, and completes only the confirmed return-or-exchange action sequence.
• Why this demonstrates the capability: The user intent is partially specified and spans multiple operations. The assistant must not guess hidden parameters or act prematurely; it must gather information interactively while keeping the task coherent and policy-compliant.

Pipeline Execution Instructions

To synthesize data for this capability, you must strictly follow a 3-phase pipeline. Do not hallucinate steps. Read the corresponding reference file for each phase sequentially:

1. Phase 1: Environment Exploration

Read the exploration guidelines to discover raw knowledge seeds: references/EXPLORATION.md

1. Phase 2: Trajectory Selection

Once Phase 1 is complete, read the selection criteria to evaluate the trajectory: references/SELECTION.md

1. Phase 3: Data Synthesis

Once a trajectory passes Phase 2, read the synthesis instructions to generate the final data: references/SYNTHESIS.md