Multi-Agent Collaborative Filtering: Orchestrating Users and Items for Agentic Recommendations

TL;DR

MACF addresses the underutilization of collaborative signals in agentic recommenders by grounding LLM-based collaboration in neighborhood evidence. It instantiates user agents for similar users and item agents for query-relevant history items, coordinated by a central orchestrator that manages multi-round discussions and personalized instructions. Empirical results across three Amazon domains show consistent improvements over strong baselines, validating the effectiveness of agentic collaborative filtering. The framework offers a scalable, explainable path toward interactive, reasoning-based recommender systems.

Abstract

Agentic recommendations cast recommenders as large language model (LLM) agents that can plan, reason, use tools, and interact with users of varying preferences in web applications. However, most existing agentic recommender systems focus on generic single-agent plan-execute workflows or multi-agent task decomposition pipelines. Without recommendation-oriented design, they often underuse the collaborative signals in the user-item interaction history, leading to unsatisfying recommendation results. To address this, we propose the Multi-Agent Collaborative Filtering (MACF) framework for agentic recommendations, drawing an analogy between traditional collaborative filtering algorithms and LLM-based multi-agent collaboration. Specifically, given a target user and query, we instantiate similar users and relevant items as LLM agents with unique profiles. Each agent is able to call retrieval tools, suggest candidate items, and interact with other agents. Different from the static preference aggregation in traditional collaborative filtering, MACF employs a central orchestrator agent to adaptively manage the collaboration between user and item agents via dynamic agent recruitment and personalized collaboration instruction. Experimental results on datasets from three different domains show the advantages of our MACF framework compared to strong agentic recommendation baselines.

Figures (2)

• Figure 1: Overview of our MACF workflow illustrated with a product recommendation example from Amazon Clothing.
• Figure 2: Distributions of MACF discussion rounds.