DeepEye: A Steerable Self-driving Data Agent System

2026-03-30

AI Summary

Existing data intelligence systems are largely confined to linear ChatBI paradigms, struggling to collaboratively manage heterogeneous data sources and often hindered by context explosion and hallucination in complex iterative analyses. This work proposes a workflow-centric autonomous data agent system that unifies structured and unstructured data through a multimodal orchestration protocol. By integrating hierarchical reasoning with context isolation mechanisms, the system decomposes intricate analytical tasks into manageable subproblems. Built upon large language models, it employs an AgentNode/ToolNode architecture, a workflow compiler, and a runtime optimizer to enable dynamic topological refinement. Experimental results demonstrate that the system generates diverse multimodal outputs—including data videos, dashboards, and analytical reports—and significantly outperforms existing approaches in execution transparency, automated optimization capability, and reliability in human-agent collaboration.

Abstract

Large Language Models (LLMs) have revolutionized natural language interaction with data. The "holy grail" of data analytics is to build autonomous Data Agents that can self-drive complex data analysis workflows. However, current implementations are still limited to linear "ChatBI" systems. These systems struggle with joint analysis across heterogeneous data sources (e.g., databases, documents, and data files) and often encounter "context explosion" in complex and iterative data analysis workflows. To address these challenges, we present DeepEye, a production-ready data agent system that adopts a workflow-centric architecture to ensure scalability and trustworthiness. DeepEye introduces a Unified Multimodal Orchestration protocol, enabling seamless integration of structured and unstructured data sources. To mitigate hallucinations, it employs Hierarchical Reasoning with context isolation, decomposing complex intents into autonomous AgentNodes and deterministic ToolNodes. Furthermore, DeepEye incorporates a database-inspired Workflow Engine (comprising a Compiler, Validator, Optimizer, and Executor) that guarantees structural correctness and accelerates execution via runtime topological optimization. In this demonstration, we showcase DeepEye's ability to orchestrate complex workflows to generate diverse multimodal outputs -- including Data Videos, Dashboards, and Analytical Reports -- highlighting its advantages in transparent execution, automated optimization, and human-in-the-loop reliability.

Problem

• Data Agents
• Heterogeneous Data Sources
• Context Explosion
• Complex Data Analysis Workflows
• Autonomous Analytics

Innovation

• Workflow-centric Architecture
• Unified Multimodal Orchestration
• Hierarchical Reasoning
• Context Isolation
• Workflow Engine