I wanted to share this research report from Enterprise Strategy Group (ESG): How Generative AI is Changing E-Commerce According to ESG, 41% of retail respondents in a 2024 survey are either already pursuing or planning to pursue AI-powered product recommendations. It’s no surprise. AI, particularly machine learning, has been a staple in eCommerce for years, especially in search, recommendation, and personalization.

In the case Vespa.ai, our AI platform has deployed machine learning (ML) models directly into its search pipelines since 2014, boosting result relevance with advanced ranking models. Vespa also offers vector search, using high-dimensional embeddings to capture semantic meaning, and supports real-time feature computation for more dynamic and context-aware ranking.

Vespa also excels at online learning and continuous model updates, enabling real-time adaptation based on user interactions—all without system downtime. By blending traditional keyword searches with ML-driven vector searches, Vespa provides a hybrid approach that balances precision with broader semantic relevance. Its capabilities for recommendations and personalization ensure that content and product suggestions are timely and aligned with user preferences. In 2022 Vespa added support for Retrieval-Augmented Generation (RAG) and so began our Generative AI journey.

But back to the ESG report, which outlines how retailers can level up their eCommerce platform with generative AI, enhancing search, recommendation, and personalization use cases and where they can expect benefits. The report also covers the complexities of scaling these applications. It’s one thing to run generative AI in the lab, but another to scale it out to support potentially millions of users or vast amounts of data.