An LLM-powered Agentic Recommendation System for Connected TV Content Discovery

2026-07-10

AI Summary

This work proposes a hybrid agent-based recommendation architecture for connected TV content discovery that effectively integrates heterogeneous contextual signals—such as trending topics and breaking news—without relying on extensive feature engineering. By orchestrating large language models (LLMs) with conventional machine learning components, the system leverages the LLM’s natural language understanding to flexibly interpret dynamic, unstructured context while preserving the efficiency of traditional retrieval pipelines. The design strategically mitigates LLM inference latency, enabling low-latency, scalable personalization. As a result, the approach maintains high recommendation accuracy while significantly enhancing responsiveness to real-time contextual cues.

Abstract

Recommendation systems, from traditional multi-stage to recent unified generative architectures, face challenges in incorporating diverse contextual signals, such as trending topics, breaking news, cultural events, and cross-surface user activities, into their ranking pipelines. These systems are designed to consume structured behavioral signals with consistent schemas, and lack the reasoning capability to naturally process unstructured or heterogeneously formatted contextual information. Incorporating such signals typically requires feature engineering, bespoke data pipelines, and carefully tuned heuristics. In this paper, we present an LLM-powered agentic recommendation system designed for Connected TV (CTV) content discovery that addresses these limitations. Our system leverages the reasoning capabilities of large language models to naturally process and synthesize diverse signals across varying schemas and structures, eliminating much of the manual integration inherent in traditional ranking and retrieval systems. Recognizing that current LLM-based solutions still fall short of traditional machine learning models in several recommendation tasks, including retrieval efficiency, personalization precision, and scalability, we adopt an agentic architecture that orchestrates specialized components, allowing each sub-task to be handled by the most suitable method, whether LLM-based or traditional ML. The main contribution of this work is our engineering approach to successfully overcoming the practical limitations of enabling LLM for recommendation, particularly inference latency. We share insights from our work and discuss the trade-offs and lessons learned in building a hybrid system that combines the flexibility of LLMs with the performance of established recommendation techniques.

Problem

• recommendation systems
• contextual signals
• large language models
• Connected TV
• unstructured data

Innovation

• LLM-powered agentic system
• contextual signal integration
• hybrid recommendation architecture
• inference latency optimization
• Connected TV content discovery