HabitatAgent: An End-to-End Multi-Agent System for Housing Consultation

Hongyang Yang; Yanxin Zhang; Yang She; Yue Xiao; Hao Wu; Yiyang Zhang; Jiapeng Hou; Rongshan Zhang

cs.LGApr 1, 2026

HabitatAgent: An End-to-End Multi-Agent System for Housing Consultation

Authors: Hongyang Yang, Yanxin Zhang, Yang She, Yue Xiao, Hao Wu, Yiyang Zhang, Jiapeng Hou, Rongshan Zhang

Organizations: Fangdongdong · Columbia University · Tsinghua University

Abstract

Housing selection is a high-stakes and largely irreversible decision problem. We study housing consultation as a decision-support interface for housing selection. Existing housing platforms and many LLM-based assistants often reduce this process to ranking or recommendation, resulting in opaque reasoning, brittle multi-constraint handling, and limited guarantees on factuality. We present HabitatAgent, the first LLM-powered multi-agent architecture for end-to-end housing consultation. HabitatAgent comprises four specialized agent roles: Memory, Retrieval, Generation, and Validation. The Memory Agent maintains multi-layer user memory through internal stages for constraint extraction, memory fusion, and verification-gated updates; the Retrieval Agent performs hybrid vector--graph retrieval (GraphRAG); the Generation Agent produces evidence-referenced recommendations and explanations; and the Validation Agent applies multi-tier verification and targeted remediation. Together, these agents provide an auditable and reliable workflow for end-to-end housing consultation. We evaluate HabitatAgent on 100 real user consultation scenarios (300 multi-turn question--answer pairs) under an end-to-end correctness protocol. A strong single-stage baseline (Dense+Rerank) achieves 75% accuracy, while HabitatAgent reaches 95%.

Figures & tables

Figure 1: HabitatAgent overview. Four specialized agents— Memory , Retrieval , Generation , and Validation —are executed as a five-stage workflow. In particular, Stage 1a (constraint extraction), Stage 1b (memory fusion), and Stage 5 (memory update) are internal substeps of the Memory Agent .

Figure 2: Use case of HabitatAgent. A user query is processed through constraint extraction, hybrid retrieval (GraphRAG), evidence-grounded generation, and multi-tier validation to produce a verified recommendation.

System	Accuracy	nDCG@5	Faithfulness	P95 Latency (ms)
B1: Monolithic RAG	0.72	0.76	0.78	450
B2: Dense+Rerank	0.75	0.80	0.82	380
B3: GraphRAG-Fixed	0.82	0.85	0.88	820
B4: LLM-Ranker	0.70	0.78	0.75	1200
B5: Self-RAG	0.78	0.82	0.85	680
B6: Rule-Verifier	0.80	0.83	0.86	520

Table 1: Overall performance comparison on the full dataset (300 queries). Best results are in bold .

System	Accuracy	CSR@5	P95 Latency (ms)
B2: Dense+Rerank	0.62	0.08	380
B3: GraphRAG-Fixed	0.85	0.88	820
Ours: HabitatAgent	0.95	0.95	680

Table 2: Performance on the complex query subset (60 queries with

\geq

3 constraints).

Configuration	Accuracy	Change
Full System (HabitatAgent)	0.95	–
w/o Verification Gate	0.88	$-7.0$ pp
w/o Adaptive Routing	0.75	$-20.0$ pp
w/o Failure Remediation	0.87	$-8.0$ pp
w/o Multi-Tier Validation	0.85	$-10.0$ pp

Table 3: Ablation study results (end-to-end accuracy).

Appendix figures & tables2 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 3: Product Demo of Fangdongdong. Scan the QR code above, or search for "Fangdongdong" in WeChat to access the product and experience the real estate decision-making system powered by HabitatAgent.

AI-Assisted Decision Making

AI-assisted decision making studies how computational systems can inform human choices while accounting for context, uncertainty, and consequences.

See all 144 papers

ReCoQA: A Benchmark for Tool-Augmented and Multi-Step Reasoning in Real Estate Question and Answering
Apr 20, 2026Yindong Zhang, Wenmian Yang, Yiquan Zhang +1LLM Agent Orchestration AI Agent Benchmarks
Hong Kong Baptist University · Beijing Normal-Hong Kong Baptist University · Beijing Normal University
AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation
Mar 4, 2026Yunxiao Shi, Wujiang Xu, Tingwei Chen +7Agent Benchmarks LLM Agents
University of Technology Sydney · Rutgers University · University of New South Wales +2
Exploring Plan Space through Conversation: An Agentic Framework for LLM-Mediated Explanations in Planning
Mar 2, 2026Guilhem Fouilhé, Rebecca Eifler, Antonin Poché +2Multi-Agent LLM Systems Explainable Artificial Intelligence
IRIT, Toulouse, France · LAAS-CNRS, Toulouse, France · IRT Saint Exupery, Toulouse, France +2

Apr 20, 2026J/K move · Enter open · S save

ReCoQA: A Benchmark for Tool-Augmented and Multi-Step Reasoning in Real Estate Question and Answering

Yindong Zhang, Wenmian Yang, Yiquan Zhang, Weijia Jia

Hong Kong Baptist University · Beijing Normal-Hong Kong Baptist University · Beijing Normal University

LLM Agent Orchestration AI Agent Benchmarks Tool-Using Agents LLM Tool Use Knowledge-Intensive QA

Open paper AskPDF arXiv

Developing agents capable of navigating fragmented, multi-source information remains challenging, primarily due to the scarcity of benchmarks reflecting hybrid workflows combining database querying with external APIs. To bridge this gap, we introduce ReCoQA, a large-scale benchmark of 29,270 real-estate instances featuring machine-verifiable supervision for intermediate steps, including structured intent labels, SQL queries, and API calls. Complementarily, we propose HIRE-Agent, a hierarchical framework instantiating an understand-plan-execute architecture as a strong baseline. By orchestrating a Front-end parser, a planning Supervisor, and execution Specialists, HIRE-Agent effectively integrates heterogeneous evidence. Extensive experiments demonstrate that HIRE-Agent constitutes a strong baseline and substantiates the necessity of hierarchical collaboration for complex, real-world reasoning tasks.

Abstract

AI-Assisted Decision Making

Explore similar work

ReCoQA: A Benchmark for Tool-Augmented and Multi-Step Reasoning in Real Estate Question and Answering

AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation

Exploring Plan Space through Conversation: An Agentic Framework for LLM-Mediated Explanations in Planning