From Experience to Skill: Multi-Agent Generative Engine Optimization via Reusable Strategy Learning

2026-04-21

AI Summary

This work addresses the limitations of existing Generative Engine Optimization (GEO) methods, which treat instances in isolation and struggle to transfer effective strategies across tasks and engines. To overcome this, the paper reframes GEO as a policy learning problem and introduces MAGEO, a multi-agent framework that enables collaborative planning, editing, and fidelity-aware evaluation. Validated editing patterns are distilled into engine-specific, reusable skills. The approach innovatively incorporates a reusable policy learning mechanism that integrates engine preference modeling with cross-task policy transfer, establishing a scalable, learning-driven GEO paradigm. Experiments on three major generative engines demonstrate that MAGEO significantly outperforms heuristic baselines, achieving consistent improvements in semantic visibility and citation fidelity. Ablation studies further confirm the critical contributions of policy reuse and engine preference modeling.

Technology Category

• Search and Optimization: Learning to Search
• Multiagent Systems: Mechanism Design
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Generative engines (GEs) are reshaping information access by replacing ranked links with citation-grounded answers, yet current Generative Engine Optimization (GEO) methods optimize each instance in isolation, unable to accumulate or transfer effective strategies across tasks and engines. We reframe GEO as a strategy learning problem and propose MAGEO, a multi-agent framework in which coordinated planning, editing, and fidelity-aware evaluation serve as the execution layer, while validated editing patterns are progressively distilled into reusable, engine-specific optimization skills. To enable controlled assessment, we introduce a Twin Branch Evaluation Protocol for causal attribution of content edits and DSV-CF, a dual-axis metric that unifies semantic visibility with attribution accuracy. We further release MSME-GEO-Bench, a multi-scenario, multi-engine benchmark grounded in real-world queries. Experiments on three mainstream engines show that MAGEO substantially outperforms heuristic baselines in both visibility and citation fidelity, with ablations confirming that engine-specific preference modeling and strategy reuse are central to these gains, suggesting a scalable learning-driven paradigm for trustworthy GEO. Code is available at https://github.com/Wu-beining/MAGEO

Problem

• Generative Engine Optimization
• strategy learning
• multi-agent
• reusable skills
• citation fidelity

Innovation

• Generative Engine Optimization
• Multi-Agent Framework
• Reusable Strategy Learning
• Citation Fidelity
• Skill Distillation