"What Are You Doing?": Effects of Intermediate Feedback from Agentic LLM In-Car Assistants During Multi-Step Processing

2026-02-17

AI Summary

This study addresses the challenge of effectively communicating reasoning processes and progress by multi-step task agents in attention-sensitive driving scenarios to optimize user experience. Through controlled experiments, it investigates how the timing and granularity of intermediate feedback provided by in-vehicle large language model agents across different task phases influence users’ cognitive load, trust, and overall experience. The work proposes an adaptive feedback strategy that initially employs high transparency to establish trust, then dynamically reduces redundancy as system reliability increases, while adjusting feedback granularity based on task criticality and contextual demands. Results demonstrate that this approach significantly enhances perceived responsiveness, trust, and user experience while reducing task load, with consistent benefits observed across varying task complexities and interaction contexts.

Technology Category

• Multiagent Systems: Agent Communication
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: User Experience and Usability

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience

Abstract

Agentic AI assistants that autonomously perform multi-step tasks raise open questions for user experience: how should such systems communicate progress and reasoning during extended operations, especially in attention-critical contexts such as driving? We investigate feedback timing and verbosity from agentic LLM-based in-car assistants through a controlled, mixed-methods study (N=45) comparing planned steps and intermediate results feedback against silent operation with final-only response. Using a dual-task paradigm with an in-car voice assistant, we found that intermediate feedback significantly improved perceived speed, trust, and user experience while reducing task load - effects that held across varying task complexities and interaction contexts. Interviews further revealed user preferences for an adaptive approach: high initial transparency to establish trust, followed by progressively reducing verbosity as systems prove reliable, with adjustments based on task stakes and situational context. We translate our empirical findings into design implications for feedback timing and verbosity in agentic assistants, balancing transparency and efficiency.

Problem

• agentic AI
• intermediate feedback
• in-car assistants
• user experience
• multi-step tasks

Innovation

• agentic LLM
• intermediate feedback
• in-car assistant
• user trust
• adaptive verbosity