CarePilot: A Multi-Agent Framework for Long-Horizon Computer Task Automation in Healthcare

TL;DR

CarePilot is a multi-agent system designed for long-horizon, complex healthcare task automation, addressing the limitations of existing models in medical workflows through iterative learning and reasoning.

Contribution

We introduce CareFlow, a comprehensive healthcare workflow benchmark, and develop CarePilot, a novel multi-agent framework that improves long-horizon reasoning in medical automation tasks.

Findings

CarePilot outperforms existing models by over 15% on benchmark tasks.

The framework effectively handles multi-step, domain-specific medical workflows.

Iterative simulation enhances the robustness and reasoning capabilities of the system.

Abstract

Multimodal agentic pipelines are transforming human-computer interaction by enabling efficient and accessible automation of complex, real-world tasks. However, recent efforts have focused on short-horizon or general-purpose applications (e.g., mobile or desktop interfaces), leaving long-horizon automation for domain-specific systems, particularly in healthcare, largely unexplored. To address this, we introduce CareFlow, a high-quality human-annotated benchmark comprising complex, long-horizon software workflows across medical annotation tools, DICOM viewers, EHR systems, and laboratory information systems. On this benchmark, existing vision-language models (VLMs) perform poorly, struggling with long-horizon reasoning and multi-step interactions in medical contexts. To overcome this, we propose CarePilot, a multi-agent framework based on the actor-critic paradigm. The Actor integrates tool grounding with dual-memory mechanisms (long-term and short-term experience) to predict the next semantic action from the visual interface and system state. The Critic evaluates each action, updates memory based on observed effects, and either executes or provides corrective feedback to refine the workflow. Through iterative agentic simulation, the Actor learns to perform more robust and reasoning-aware predictions during inference. Our experiments show that CarePilot achieves state-of-the-art performance, outperforming strong closed-source and open-source multimodal baselines by approximately 15.26% and 3.38%, respectively, on our benchmark and out-of-distribution dataset.