Bootstrapping Conversational Recommendation Agents At Spotify: Synthetic Data Generation and Self-Improvement Loops

Spotify productionized a pipeline for multi-turn synthetic data generation and a self-improvement loop that improved conversational recommendation agent planning quality by +8% over a highly optimized manual prompt, according to an arXiv paper (2609.30297v1). The self-improvement loop combines variance-based contrastive optimization with iterative refinement through a coding agent to automatically identify and fix planning and tool-use errors. Online A/B tests showed +14% user listening, +5% increase in weekly active users, and a 5% reduction in skip rate versus a prior experience supporting only session refinement.

arXiv:2609.30297v1 Announce Type: new Abstract: Conversational recommendation agents are a new paradigm for content discovery, enabling users to express complex intents through natural language e.g., "recommend Italian indie artists I haven't heard before" . A central challenge in building such agents is optimizing agent planning -- deciding how to select, sequence, and invoke tools -- particularly in cold-start settings where real user interactions are not yet available. We introduce a pipeline for multi-turn synthetic data generation and a self-improvement loop to address this challenge. The synthetic data pipeline transforms single-turn prompts into realistic multi-turn conversations, enabling systematic evaluation before launch. The self-improvement loop combines variance-based contrastive optimization with iterative refinement through a coding agent, automatically identifying and fixing planning and tool-use errors. Our approach improves quality by +8% on top of a highly optimized manual prompt. The system has been productionized and significantly accelerated iteration cycles for the launch of a conversational recommendation agent at Spotify. Online A/B tests demonstrate its effectiveness, with +14% user listening, +5% increase in weekly active users, and a 5% reduction in skip rate compared to a prior experience supporting only session refinement. This work provides a practical framework for accelerating the development of conversational recommendation agents in industry.