BizSage: A Self-Evolving Multi-Agent Framework for Business Research with Efficient Knowledge Retrieval

Yuhe Wu; Guangyu Wang; Jiaxin Liu; Guang Zhang

cs.CLSep 3, 2026

BizSage: A Self-Evolving Multi-Agent Framework for Business Research with Efficient Knowledge Retrieval

Authors: Yuhe Wu, Guangyu Wang, Jiaxin Liu, Guang Zhang

Organizations: The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China

Abstract

While multi-agent systems based on large language models (LLMs) have shown promise in automating the progressive workflow of academic research, extending them to economics and business research, where specialized domain knowledge spans neighboring disciplines yet remains difficult to access in a structured way, presents two challenges. First, existing methods mostly retrieve at the paper level, yet the evidence needed for research tasks is often distributed across different sections, creating a granularity mismatch that hinders retrieval coverage and precision. Second, these fields demand strict empirical rigor, yet current systems provide limited mechanisms for learning from evaluation feedback. We present BizSage, a multi-agent framework combining corpus-level fine-grained retrieval with quality-driven self-evolution. We build a Lateral Knowledge Graph (LKG) by merging section-level knowledge graphs and apply Personalized PageRank (PPR) to surface semantically relevant and structurally important sections. Seven specialized agents collaborate under a Meta-Review self-evolution mechanism that distills failure modes from evaluation traces into reusable strategies. On a benchmark spanning four domains and three tasks, BizSage ranks first on the majority of metrics, achieves pairwise win-rates above 60% against six baselines, and produces zero hallucinated citations. We hope BizSage paves the way for reliable research assistance in economics, business, and the broader social sciences.

Figures & tables

Figure 1: Comparison of three retrieval methods. Left : online API search returns selected information but is slow. Middle : offline database retrieves whole papers but includes irrelevant content. Right : our approach builds a Lateral Knowledge Graph and retrieves selected sections, achieving both speed and accuracy.

Figure 2: Overview of the proposed framework. The left part illustrates Stage 1 (literature retrieval) and the right part illustrates Stage 2 (grounded response generation).

	Survey				Idea Formulation				Research Plan
Method	Rel.	Grd.	Cov.	Syn.	Rel.	Grd.	Nov.	Fea.	Rel.	Grd.	Rig.	Fea.
GPT-5.5
AI-Scientist-v2	4.68	3.57	3.97	3.94	4.68	3.29	3.69	3.85	4.88	4.22	4.05	3.84
EvoScientist	4.55	2.20	3.97	3.82	4.72	2.28	2.53	3.75	4.43	1.84	2.80	2.76
InternAgent-1.5	3.52	1.97	2.54	2.84	4.19	2.16	3.04	3.36	4.22	1.56	3.87	3.03
ResearchAgent	4.37	2.47	3.81	3.82	4.33	2.87	3.32	3.66	4.83	2.55	3.17	3.52

Table 1: Main results across three backbone LLMs. Bold and underline mark the best and second-best scores in each column within each backbone.

\dagger

denotes task-specific systems evaluated only on their target task. Metric abbreviations: Rel. = Relevance, Grd. = Grounding, Cov. = Coverage, Syn. = Synthesis, Nov. = Novelty, Fea. = Feasibility, Rig. = Rigor.

Figure 3: Pairwise win-rate (%) of BizSage against each baseline, judged by GPT-4o. Each sub-figure shows one task type, with bars grouped by backbone.

	Survey				Idea Formulation				Research Plan
Method	Rel.	Grd.	Cov.	Syn.	Rel.	Grd.	Nov.	Fea.	Rel.	Grd.	Rig.	Fea.
GPT-5.5
BizSage (Iteration 1)	4.14	3.05	3.77	3.86	4.52	2.07	3.16	3.54	4.60	3.35	3.94	3.31
BizSage (Iteration 2)	4.30	4.06	3.94	3.89	4.57	3.17	3.40	3.67	4.62	3.75	4.00	3.42
BizSage (Iteration 3)	4.77	4.14	4.04	3.99	4.71	3.42	3.43	3.95	4.89	4.36	4.07	3.95
Gemini-3-Flash

Table 2: Self-evolution across three strategy-refinement iterations per backbone. Each iteration refines strategies using 10 development queries. Bold and underline mark the best and second-best scores per column within each backbone.

	Survey				Idea Formulation				Research Plan
Method	Rel.	Grd.	Cov.	Syn.	Rel.	Grd.	Nov.	Fea.	Rel.	Grd.	Rig.	Fea.
BizSage (full)	4.77	4.14	4.04	3.99	4.71	3.42	3.43	3.95	4.89	4.36	4.07	3.95
Ablations
$\vdash$ without Section-Level Knowledge Graph	4.32	3.19	3.99	3.92	4.48	2.26	3.35	3.61	4.73	3.56	4.01	3.55
$\vdash$ without Reviewer	4.16	4.06	3.89	3.83	4.36	3.08	3.35	3.50	4.62	3.27	3.97	3.32
$\vdash$ without Discussant	4.20	4.03	3.79	3.68	4.26	3.07	3.19	3.36	4.57	3.26	3.87	3.18

Table 3: Ablation study on the GPT-5.5 backbone. Each variant removes one component from the full system. Bold and underline mark the best and second-best scores per column.

Figure 4: Efficiency–performance matrix across the three task types. Each point represents one method; the x-axis shows mean quality score and the y-axis shows retrieval time in seconds. Methods closer to the lower-right corner achieve higher quality with shorter retrieval time.

Extraction Model	NV	RS	MC
Qwen3-8B	65.0	44.4	42.9
GPT-4o	90.0	88.2	92.3
GPT-5.5	95.2	94.4	93.3

Table 4: Quality of section-level graph extraction across extraction models on 50 manually annotated paper sections. NV = Node Validity, RS = Relation Support, MC = Merge Consistency. All values are percentages, and higher is better.

Method	#Citations	Hallucination Rate
BizSage (Ours)	894	0.0%
InternAgent-1.5	98	5.1%
EvoScientist	75	2.7%
AI-Scientist-v2	0	–
ResearchAgent	0	–
Idea2Story ∗	0	–

Table 5: Citation integrity on the survey task. ∗ These systems do not produce surveys; we evaluate twenty outputs from their native tasks instead.

Appendix figures & tables20 assets

Supplementary material from the paper’s appendix.

Appendix

Journal	Publisher	Coverage	Website	Papers
American Economic Review	AEA	1980–2021	aeaweb.org/j/aer	8,760
Journal of Econometrics	Elsevier	1973–2022	elsevier.com/j/econometrics	5,637
Journal of Economic Theory	Elsevier	1974–2022	elsevier.com/j/econ-theory	5,151
The Economic Journal	OUP	1927–2023	oup.com/ej	4,832
Econometrica	Econometric Soc.	1980–2023	econometricsociety.org	4,622
Review of Economics and Statistics	MIT Press	1969–2018	direct.mit.edu/rest	3,744

Appendix

Table 6: Economics journals in the source corpus.

Journal	Publisher	Coverage	Website	Papers
Journal of Finance	Wiley	1949–2020	wiley.com/j/15406261	5,902
Journal of Financial Economics	Elsevier	1978–2022	elsevier.com/j/jfe	3,201
Review of Financial Studies	OUP	1988–2021	oup.com/rfs	2,529
Journal of International Money and Finance	Elsevier	1982–2023	elsevier.com/j/jimf	2,264
Journal of Corporate Finance	Elsevier	1997–2021	elsevier.com/j/jcf	2,214
Insurance: Mathematics and Economics	Elsevier	1982–2022	elsevier.com/j/ime	1,806

Appendix

Table 7: Finance journals in the source corpus.

Journal	Publisher	Coverage	Website	Papers
INFORMS Journal on Applied Analytics	INFORMS	1971–2023	informs.org/j/inte	3,204
Operations Research	INFORMS	1976–2024	informs.org/j/opre	2,164
Transportation Science	INFORMS	1967–2011	informs.org/j/trsc	1,328
Management Science	INFORMS	1970–1977	informs.org/j/mnsc	1,043
Mathematical Programming	Springer	1995–2024	springer.com/j/10107	70
Operations Research Letters	Elsevier	2004–2023	elsevier.com/j/orl	38

Appendix

Table 8: Operations management journals in the source corpus.

Journal	Publisher	Coverage	Website	Papers
Biometrika	OUP	1925–2019	oup.com/biomet	6,947
Computational Statistics & Data Analysis	Elsevier	1990–2022	elsevier.com/j/csda	4,906
Journal of Multivariate Analysis	Elsevier	1976–2022	elsevier.com/j/jmva	4,791
Journal of Business & Economic Statistics	Taylor & Francis	1983–2024	tandfonline.com/ubes20	3,418
JRSS Series C (Applied Statistics)	OUP	1950–2019	oup.com/jrsssc	3,411
JRSS Series B (Statistical Methodology)	OUP	1936–2010	oup.com/jrsssb	1,717

Appendix

Table 9: Statistics journals in the source corpus.

A. Relevance Survey, Idea Formulation, Research Plan
1	Output does not address the query topic.
2	Partially addresses the query but drifts substantially.
3	Addresses the query topic with noticeable tangential content.
4	Stays on topic with only minor digressions.
5	Directly and fully addresses the query.
B. Coverage Survey

Appendix

Table 10: Scoring rubrics for the seven evaluation dimensions. Each is rated 1–5.

Task	Instances	Unique	1	2	3	4	5+
Survey	4,121	3,710	3,384	262	46	16	2
Idea Formulation	3,039	2,725	2,449	245	26	4	1
Research Plan	2,151	1,936	1,757	149	26	2	2
Overall	9,311	7,535	6,269	951	204	68	43

Appendix

Table 11: Citation-frequency distribution over all 300 outputs. The numbered columns give how many unique papers appear in exactly that many outputs.

Rank	Most-Cited Paper	Domain	Main Task(s)	Outputs	Share
1	Are Trading Imbalances Indicative of Private Information?	Finance	All three	9	3.0%
2	Network Design and Transportation Planning: Models and Algorithms	Operations Mgmt.	All three	9	3.0%
3	Enhanced Branch and Price and Cut for Vehicle Routing with Split Deliveries and Time Windows	Operations Mgmt.	All three	8	2.7%
4	Price Discovery and Liquidity Recovery: Forex Market Reactions to Macro Announcements	Finance	Survey, Idea Form.	7	2.3%
5	Robustness of the Linear Mixed Model to Misspecified Error Distribution	Statistics	All three	7	2.3%
6	Branch-and-Cut-and-Price for the Pickup and Delivery Problem with Time Windows	Operations Mgmt.	All three	7	2.3%

Appendix

Table 12: Ten most frequently cited papers. Outputs is the number of outputs citing the paper and Share is the corresponding fraction of the 300 outputs.

	Survey	Idea Form.	Res. Plan	Total
Economics	261	389	350	1000
Finance	228	364	408	1000
Operations Mgmt.	196	427	377	1000
Statistics	243	398	359	1000
Avg. $\sigma$	0.76	0.83	0.80	0.80

Appendix

Table 13: Classification distribution of 4,000 held-out papers. Idea formulation and research plan papers outnumber survey papers across all domains; confidence is highest for idea formulation.

	Survey		Idea Form.		Res. Plan		Total
	Ev	Dev	Ev	Dev	Ev	Dev	Ev	Dev
Economics	29	26	24	27	27	24	80	77
Finance	24	27	28	25	30	33	82	85
Operations Mgmt.	22	25	26	23	23	21	71	69
Statistics	25	22	22	25	20	22	67	69
Total	100	100	100	100	100	100	300	300

Appendix

Table 14: Distribution of the 600 benchmark queries. Ev = evaluation set, Dev = development set for self-evolution.

Dimension	Human ICC	LLM–Human $\rho$
Relevance	0.85	0.89
Grounding	0.82	0.86
Coverage	0.79	0.83
Synthesis	0.74	0.78
Novelty	0.71	0.75
Feasibility	0.76	0.80

Appendix

Table 15: LLM judge validation. Human ICC: intraclass correlation among five annotators. LLM–Human

\rho

: Spearman correlation between GPT-4o scores and averaged human scores. All correlations are significant at

p<0.01

Method	Survey	Idea Form.	Res. Plan
LKG (no diffusion)	$3.806\pm 0.775$	$3.828\pm 0.889$	$3.283\pm 0.853$
Long-Walk PPR	$6.991\pm 1.313$	$6.327\pm 1.569$	$6.689\pm 1.318$
Katz Diffusion	$4.404\pm 2.092$	$4.485\pm 3.640$	$3.542\pm 0.994$
Heat-Kernel Diffusion	$5.752\pm 2.133$	$7.677\pm 2.596$	$4.666\pm 1.584$
AI Co-Scientist	–	–	$2.176\pm 0.666$
Idea2Story	–	$2.602\pm 1.595$	–

Appendix

Table 16: Runtime efficiency across task types. All runtimes are reported in seconds.

Autonomous Research Agents

Autonomous research agents use language models and tools to search literature, design and run experiments, analyze results, and support scientific discovery.

See all 326 papers

Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework
Apr 7, 2026Komal Kumar, Aman Chadha, Salman Khan +2AI-Assisted Scientific Research Information Retrieval
Mohamed bin Zayed University of Artificial Intelligence · AWS Generative AI Innovation Center, Amazon Web Services
HLER: Human-in-the-Loop Economic Research via Multi-Agent Pipelines for Empirical Discovery
Mar 8, 2026Chen Zhu, Xiaolu WangMulti-Agent LLM Systems Human-in-the-Loop AI
China Agricultural University
Agent-Orchestrated Adaptive RAG: A Comparative Study on Structured and Multi-Hop Retrieval
Jun 4, 2026Anuj Maharjan, Devinder Kaur, Richard MolyetRetrieval-Augmented Generation Multi-Hop Retrieval
Dept. of Electrical Engineering and Computer Science University of Toledo Toledo, OH, USA

Apr 7, 2026J/K move · Enter open · S save

Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework

Komal Kumar, Aman Chadha, Salman Khan, Fahad Shahbaz Khan+1

Mohamed bin Zayed University of Artificial Intelligence · AWS Generative AI Innovation Center, Amazon Web Services

AI-Assisted Scientific Research Information Retrieval Multi-Agent LLM Systems

Open paper AskPDF arXiv

The rapid growth of scientific literature has made it increasingly difficult for researchers to efficiently discover, evaluate, and synthesize relevant work. Recent advances in multi-agent large language models (LLMs) have demonstrated strong potential for understanding user intent and are being trained to utilize various tools. In this paper, we introduce Paper Circle, a multi-agent research discovery and analysis system designed to reduce the effort required to find, assess, organize, and understand academic literature. The system comprises two complementary pipelines: (1) a Discovery Pipeline that integrates offline and online retrieval from multiple sources, multi-criteria scoring, diversity-aware ranking, and structured outputs; and (2) an Analysis Pipeline that transforms individual papers into structured knowledge graphs with typed nodes such as concepts, methods, experiments, and figures, enabling graph-aware question answering and coverage verification. Both pipelines are implemented within a coder LLM-based multi-agent orchestration framework and produce fully reproducible, synchronized outputs including JSON, CSV, BibTeX, Markdown, and HTML at each agent step. This paper describes the system architecture, agent roles, retrieval and scoring methods, knowledge graph schema, and evaluation interfaces that together form the Paper Circle research workflow. We benchmark Paper Circle on both paper retrieval and paper review generation, reporting hit rate, MRR, and Recall at K. Results show consistent improvements with stronger agent models. We have publicly released the website at https://papercircle.vercel.app/ and the code at https://github.com/MAXNORM8650/papercircle.

Abstract

Autonomous Research Agents

Explore similar work

Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework

HLER: Human-in-the-Loop Economic Research via Multi-Agent Pipelines for Empirical Discovery

Agent-Orchestrated Adaptive RAG: A Comparative Study on Structured and Multi-Hop Retrieval