Organizations: The Hong Kong University of Science and Technology (Guangzhou), China · Shanghai Jiao Tong University, Shanghai, China · Renmin University of China, Beijing, China · Beijing Institute of Technology, Beijing, China · Huawei · Southeast University, Nanjing, China · Tsinghua University, Beijing, China · DeepWisdom
The rapid advancement of large language models (LLMs) has spurred the emergence of data agents, autonomous systems designed to orchestrate Data + AI ecosystems for tackling complex data-related tasks. However, the term "data agent" currently suffers from terminological ambiguity and inconsistent adoption, conflating simple query responders with sophisticated autonomous architectures. This terminological ambiguity fosters mismatched user expectations, accountability challenges, and barriers to industry growth. Inspired by the SAE J3016 standard for driving automation, this survey introduces the first systematic hierarchical taxonomy for data agents, comprising six levels that delineate and trace progressive shifts in autonomy, from manual operations (L0) to a vision of generative, fully autonomous data agents (L5), thereby clarifying capability boundaries and responsibility allocation. Through this lens, we offer a structured review of existing research arranged by increasing autonomy, encompassing specialized data agents for data management, preparation, and analysis, alongside emerging efforts toward versatile, comprehensive systems with enhanced autonomy. We further analyze critical evolutionary leaps and technical gaps for advancing data agents, especially the ongoing L2-to-L3 transition, where data agents evolve from procedural execution to autonomous orchestration. Finally, we conclude with a forward-looking roadmap, envisioning proactive, generative data agents. A curated companion repository is available at https://github.com/HKUSTDial/awesome-data-agents.
Figures & tables
Aspect
General-purpose LLM Agents
Data Agents
Core Goal
Broad User Tasks: Assist with writing, search, planning, dialogue, and content generation.
Data-centric Tasks: Support data management, preparation, analysis, operation, and governance.
Operating Environment
Open-ended Digital Environments: Interact with task-dependent contexts, tools, webpages, or APIs.
Data Infrastructure: Operate over DBMSs, data lakes, BI systems, ETL tools, and runtimes.
Grounding and Interaction
Prompt- and Tool-grounded: Reason over prompts, retrieved content, and tool observations when available.
Data-grounded and Execution-aware: Ground reasoning in schemas, values, metadata, constraints, provenance, and execution feedback.
Tool and Execution
General-purpose Tool Use: Invoke tools such as search engines, calculators, OCR, code, or web APIs.
Specialized Data Tool Use: Invoke SQL engines, DB loaders, profilers, optimizers, validators, visualization libraries, or code/runtime engines.
Outputs and Evidence
User-facing Artifacts: Generate text, plans, content, answers, or tool-mediated actions.
Verifiable Data Artifacts: Produce SQL, configurations, cleaned/integrated data, charts, reports, workflows, or execution traces.
Reliability and Evaluation
Direct Task Reliability: Emphasize response correctness, task success, and tool-use robustness.
Downstream Workflow Reliability: Require data grounding, workflow robustness, auditability, and control of cascading errors.
Table 1: Comparison between General-purpose LLM Agents and Data Agents
Figure 2: Representative Data Agents Across Different Levels.
Figure 3: Evolutionary Leaps Between Data Agent Levels.
Figure 4: L1 Data Agents (Assistance).
Task
Data Agent
In-Context Learning
RAG
SFT
RL
Data Complexity
Multis.
Hete.
Multim.
Data Management
Configuration Tuning
LLMTune [ 44 ]
zero-shot
-
✓
-
-
✓
-
GPTuner [ 45 ]
zero-shot
-
-
-
✓
✓
-
λ -Tune [ 46 ]
zero-shot
✓
-
-
-
-
-
E2ETune [ 47 ]
few-shot
-
✓
-
-
-
-
Query Optimization
DB-GPT [ 48 ]
few-shot
✓
✓
-
✓
✓
-
Table 2: Comparison of Representative L1 Data Agents. RAG: Retrieval-Augmented Generation; SFT: Supervised Fine-Tuning; RL: Reinforcement Learning. Data complexity: Multi-source (Multis.), Heterogeneous (Hete.), and Multimodal (Multim.).
Table 4: Comparison of Representative Proto-L3 Data Agents from Academia Research and Industry Products. Compares Open-source: availability; Undef Ops.: capabilities in utilizing unpredefined operators; data-related task coverage across data management, preparation, analysis; data complexity dimensions: Multi-source (Multis.), Heterogeneous (Hete.), and Multimodal (Multim.). † indicates industry system.
Traditional data systems face profound limitations in the AI era, relying on human-crafted pipelines, lacking semantic understanding of heterogeneous data, and operating through rigid, reactive processing. To address these challenges, we propose a new paradigm called the Data Agent, designed to manage, process, and analyze data with minimal human intervention. Data agents autonomously execute a wide range of data-related tasks, transforming traditional data systems by shifting from manual design to autonomous orchestration, from literal manipulation to semantic interpretation, and from reactive to proactive processing. Our Data Agent system includes six components: semantic data organization, semantic operators, agentic pipeline orchestration and optimization, feedback-driven refinement, memory management, and proactive adaptation. Building on this foundation, we also develop two specialized agents: the data analytics agent and the data science agent. Experiments on real benchmarks demonstrate significant performance gains of our data agent over state-of-the-art methods. We identify open challenges to guide future research in building fully autonomous data systems.
Guoliang Li, Peiyao Zhou, Xuanhe Zhou +3
Department of Computer Science, Tsinghua University, Beijing, China · Shanghai Jiao Tong University, Shanghai, China · The Hong Kong University of Science and Technology (Guangzhou), China +1
Answering questions over heterogeneous data means combining evidence from files, databases, documents and the web. Data agents automate this, but their success depends on context. Three context-engineering techniques (CETs) help by shaping it: semantic layers describe the data, orchestration splits work among agents, and compaction condenses interaction history. Existing studies validate each CET alone, leaving open how agent behavior shapes its effect. We present a comprehensive analysis of the three CETs, alone and combined, with seven large language models (LLMs) on 1,514 long-horizon tasks from five benchmarks, covering behavior, accuracy and cost. We find that data agents without CETs form three behavior groups: investigators, premature answerers that answer without inspecting data, and early stoppers that give up. Each CET delivers its known benefit, from more investigation to smaller prompts, but only the behavior group it matches turns that benefit into accuracy, such as premature answerers under orchestration. For investigators, the same CETs can hurt: the semantic layer stops them from looking, orchestration makes them overconfident, costing 5 percentage points, and compaction makes them repeat steps. Combining CETs yields smaller gains and losses than their sum. We derive guidance for choosing CETs by agent behavior, task demands and accuracy-cost trade-offs.
Zhivar Sourati, Marianne Menglin Liu, Zheng Wang +6
University of Southern California · Oracle AI · University of Pennsylvania
Large language model agents are increasingly deployed for data-intensive work, yet reliable data analysis requires more than general-purpose reasoning and ad hoc tool augmentation. Data Agents, equipped with workflow harnesses, offer a promising paradigm for automating the end-to-end data science lifecycle. This paper examines Data Agents from a harness-centric perspective. First, we introduce a taxonomy of Data Agents and associated data environments, organizing the literature around five functional stages: perception, planning, execution, verification, and repair. Second, we analyze the key technical routes within each stage, identifying 15 distinct approaches ranging from data structure probing to data state reconstruction. Third, we identify four open reliability problems: inactive semantic calibration, missing clarification, missing experience transfer, and the missing verification-repair repository. These problems explain why silent failures can persist even when individual components function correctly, highlighting the need for rigorous workflow harnesses and shared reliability resources. Finally, we summarize the horizontal task families of Data Agents, examine their vertical application settings, and benchmarks for evaluation, while maintaining a companion repository at https://github.com/DEEP-PolyU/Awesome-Data-Agents.
Huachi Zhou, Yujing Zhang, Jiahe Du +7
The Hong Kong Polytechnic University, Hong Kong, China · Jilin University, Changchun, China