How WP Engine Uses Elastic and Gemini to Power AI Search for 1.5M WordPress Users

WP Engine has been the backbone of the WordPress ecosystem since 2010, providing managed hosting and developer tools to more than 1.5 million users across 200,000+ websites spanning 150 countries. As AI capabilities moved from experiments to customer expectations, WP Engine found itself at a crossroads: its clients — ranging from media companies to e-commerce brands — needed advanced AI features on their websites, but building those features across the sprawling and inconsistent landscape of WordPress configurations required a technology partner that could handle the complexity at scale.

The challenge was structural. WordPress data exists in countless configurations: different themes, plugins, custom post types, and content structures. Building consistent AI search or recommendation capabilities across this variety demanded a platform that could extract any data type, normalize it, and store it as vectors — all while maintaining the speed and reliability that web visitors expect. WP Engine’s existing infrastructure was not built for vector search or generative AI, and developing those capabilities from scratch would have consumed engineering resources that should be directed at product innovation.

WP Engine chose Elastic’s Search AI Platform as the foundation, deploying Elasticsearch’s vector database at the heart of its technology stack and integrating it with Google Cloud Vertex AI. This pairing gave WP Engine a natively integrated retrieval engine that combines the multimodal strengths of Google’s Gemini models with Elastic’s AI-powered semantic and hybrid search. The practical output was Smart Search AI — a WP Engine product feature that delivers natural language search results for any WordPress site. The same infrastructure enables RAG applications: WP Engine customers can now build AI chatbots and recommendation engines grounded in their own business data without maintaining custom pipelines.

The reliability test came when a large media customer used WP Engine’s search API in a way that sent traffic on the production server from hundreds of thousands to tens of millions of queries within minutes. The result: zero downtime. The Elastic Cloud infrastructure absorbed the spike without a single disruption. Response times across the platform stabilized at around five milliseconds — fast enough to deliver seamless real-time experiences for website visitors and protect conversion rates. “There was zero downtime, not a single hiccup,” said Shane Daly, Principal Software Engineer at WP Engine. “The Elastic Cloud infrastructure scaled up exactly as we had hoped.”

For WP Engine and its customers, the Elastic and Google Cloud integration has shifted AI deployment from a multi-month engineering project to a rapid capability rollout. The same media customer that triggered the traffic spike was previously using a custom recommendation engine integrated with a separate LLM; that same capability is now available natively through WP Engine’s platform, eliminating the need for a bespoke solution. As AI search and personalization become baseline expectations for website visitors, WP Engine is positioned as the platform that makes those capabilities accessible to organizations of any size.