From Raw IDs to Semantic Planning: How Recommender Systems Utilize Information at Scale

2026-07-10

AI Summary

Traditional recommender systems rely on semantically opaque raw IDs, which hinder deep integration of content and contextual information for meaningful understanding. This work proposes a novel paradigm—semantic planning—that first predicts the user’s next interaction in terms of high-level semantic intent and then instantiates it into specific items or generated content. By integrating large-scale industrial architectures, multimodal signals, cross-domain modeling, and generative AI, the study explores practical pathways for implementing semantic IDs and planning mechanisms. It systematically articulates the generational evolution of recommender systems from raw IDs to semantic IDs and, for the first time, positions semantic planning as the cornerstone of next-generation recommendation. The paper provides a theoretical framework and technical roadmap toward building recommender systems endowed with semantic comprehension and proactive planning capabilities.

Abstract

The evolution of recommender systems can be explored by asking how they utilize information at scale. Throughout most of the historical period under consideration during the past two decades, industrial systems have relied on raw IDs, which are discrete, globally unique, and semantically opaque identifiers that enable exact lookup, logging, and item-specific memorization at scale. Over time, however, recommender systems have sought to utilize richer sources of information, including item content, context, multimodal signals, and cross-domain structure. This development has led to a new stage in which part of such information is no longer used solely as auxiliary features around item identity, but is increasingly encapsulated in semantic IDs that provide a more structured, model-facing form of identity. We argue that this shift goes beyond the rise of generative recommendation over traditional methods. Indeed, it reflects a broader evolution in how recommender systems utilize information under industrial-scale constraints. This paper looks at the past, present, and future to examine three connected questions: why raw IDs dominated the early development of recommender systems, why semantic information is increasingly being encapsulated in IDs today, and what may come next once recommendations move beyond semantic retrieval. In particular, we introduce semantic planning as a possible future direction in which the system first predicts the semantic target of the next exposure, and only then instantiates that target as a specific item or generated creative. We further argue that such a shift may require changes not only in model design but also in evaluation and in the way recommender systems coordinate the objectives of users, platforms, and providers.

Problem

• recommender systems
• semantic planning
• raw IDs
• semantic IDs
• information utilization

Innovation

• semantic planning
• semantic IDs
• generative recommendation
• information utilization
• recommender system evolution