#USA #Redwood City #AI Data #Zilliz #Milvus

Zilliz Predicts a New Era for Data Interaction

In a significant forecast presented on November 14, 2025, Zilliz, the innovative company behind Milvus—the leading open-source vector database—has stated that by 2026, enterprises will predominantly utilize natural language interfaces for interacting with their databases. This leap indicates a vital transformation in how organizations manage and query their data, marking a shift from traditional SQL frameworks to more intuitive and accessible natural language processing (NLP) models.

James Luan, Zilliz's VP of Engineering, expressed that we are on the verge of an era where communicating with databases will outpace traditional scripting methods in productivity. He stated, "SQL will still matter, but it no longer defines how people interact with data. Natural language is becoming the default interface for AI workloads because it lets people focus on intent instead of syntax."

The Rise of Conversational Data Access

Natural language interfaces provide users with the ability to articulate their data inquiries in everyday language, such as asking, "Show me customers whose behavior changed most in the last 30 days." In response, AI agents within these systems will adeptly translate these queries into actionable execution plans, which may include structured filtering, vector similarity searches, or hybrid queries. This paradigm shift reduces the dependency on specialists who usually translate business requirements into executable SQL, thereby enhancing accessibility for various teams including product development, analytics, and business operations.

The Limitations of SQL in Modern Data Management

SQL's design limitations become apparent in the context of advanced AI applications, which often work with vector embeddings—complex semantic representations of text, images, and other multimodal content. Such workloads cannot be efficiently expressed or managed through traditional relational queries, which is where Milvus shines.

Benchmark tests conducted by Zilliz demonstrated that Milvus significantly outperforms PostgreSQL configured with pgvector, achieving latency reductions of 60% and an impressive 4.5-fold increase in throughput under identical vector search conditions. This advantage is crucial as enterprises begin managing billions of embeddings that require rapid query responses.

Luan noted, "SQL was never designed for similarity search across thousands of dimensions. AI workloads require a semantic retrieval layer, not a relational one." The shift to natural language interfaces is not just a trend but an essential evolution in data interaction necessary to fulfill the demands of contemporary AI applications.

Accelerated Adoption Across Enterprises

The momentum behind this shift is reflected in the growing adoption of Zilliz solutions, with more than 10,000 organizations worldwide utilizing Milvus and its managed counterpart, Zilliz Cloud, for a multitude of AI applications including semantic search, recommendation engines, and retrieval-augmented generation (RAG). These implementations routinely handle operations at scales of billions of vectors while maintaining sub-10 millisecond query latency across major cloud platforms such as AWS, Google Cloud, and Microsoft Azure.

For a deeper understanding of this trend, Zilliz has shared comprehensive insights in their blog post titled Why AI Databases Don't Need SQL. Organizations aiming to explore the use of natural language interfaces for data retrieval can visit zilliz.com for more information.

About Zilliz

Founded with the goal of revolutionizing data management for AI, Zilliz continues to build the highest-performing and cost-effective vector database powered by the open-source platform Milvus. The company provides unmatched support for unstructured data on a grand scale, capable of managing billion-vector workloads while ensuring fast hybrid searches with low latency. With over 10,000 organizations harnessing Zilliz for their GenAI applications, the firm exemplifies the future of data interaction in the world of AI.