Guest:  Bianca Lewis (LinkedIn)
Company:  OpenSearch Software Foundation
Show Name: 2026 Predictions
Topic: AI Infrastructure

Vector search is now table stakes—every database offers it in some form. But as enterprises move beyond simple vector capabilities into agentic AI deployment, the infrastructure requirements are about to shift dramatically. Bianca Lewis, Executive Director of the OpenSearch Software Foundation, shares predictions for how AI agents, data quality over data size, and open platforms will define enterprise technology in 2026.

From Vector Search to Agent-Ready Databases

The database landscape underwent a dramatic transformation in 2025, with vector search capabilities becoming universal across both modern and legacy platforms. “I was looking a few days ago at the 2025 database landscape, and the one pattern that came through over and over again was that every database today is a vector database and a vector search database,” Lewis observes. Even older, established databases bolted on vector functionality to remain competitive.

This ubiquity means vector search alone no longer differentiates platforms. Lewis predicts 2026 will be the year of agent-ready databases—systems designed specifically to support AI agents that can access and interact with multiple external systems. The distinction is significant: traditional vector search might return comfortable shoes based on a query, but an agent-ready database enables AI to access Chrome browsing history, past purchase records, and return data to provide contextually relevant recommendations based on actual user behavior.

The Foundation That Won’t Go Away

Despite the rush toward AI agents and hybrid search capabilities, Lewis emphasizes that traditional keyword search remains foundational. “The more things change, the more things will be the same,” she notes. Drawing an analogy to the evolution of transportation, Lewis explains that just as wheels remained essential from bicycles to space shuttles, keyword search will continue to be critical for accurate results and scalability.

“When we invented the wheel, and then many years later developed the bicycle, the automobile, the airplane, and the space shuttle, those wheels remained foundational,” Lewis says. The nature of search has evolved, but lexical keyword search provides the accuracy and scale that enterprises need, even as AI capabilities expand around it.

Strategic AI Adoption Replaces Ad Hoc Experimentation

The unstructured, team-by-team approach to AI adoption that characterized recent years is giving way to more strategic implementation. “Companies were using AI in a very unstructured way up until now. Teams were using it individually. We know it increases productivity, but it was less organized, it was less systematic,” Lewis explains.

In 2026, enterprises will shift toward systematic AI deployment designed to drive business outcomes rather than simply boost individual contributor productivity. This strategic approach will make AI implementations more predictable and allow organizations to properly plan resources and measure impact.

Data Quality Trumps LLM Size

One of Lewis’s most pointed predictions challenges the assumption that larger language models deliver better results. “When it comes to LLMs, bigger is not always better,” she states. “It’s not about the size of the data; it’s about data quality, the context of that data, and our ability to trust the results it returns.”

The focus must shift to the data layer accessible to LLMs rather than model size alone. Technologies like retrieval augmented generation (RAG) are redefining the quality of insights by minimizing hallucinations and ensuring that AI systems return trustworthy information grounded in verified data sources.

Infrastructure Challenges at the Speed of Innovation

The rapid advancement of agentic AI will force organizations to play catch-up on infrastructure capable of supporting this new pace of innovation. Lewis identifies this as a critical challenge: developing systems that can not only support AI agents but also keep costs reasonable while maintaining performance.

For OpenSearch users specifically, the challenge extends to maximizing value from existing data. “OpenSearch is a Swiss Army knife of data platforms that we can use for multiple different use cases,” Lewis explains. Organizations often deploy OpenSearch for observability but fail to recognize that the same data can power enterprise search, AI applications, and other use cases—creating opportunities for greater efficiency.

The Open Platform Imperative

Lewis’s strongest advice for 2026 centers on maintaining open platforms. “Ensure that your platforms remain open,” she emphasizes. “The one thing we do know is that we are entering a pace of innovation that has never been seen before—not in 2025, 2024, or at any other time.”

Vendor lock-in becomes increasingly risky as innovation accelerates. Open platforms provide the control, flexibility, and predictability enterprises need to align technology with business goals and scale appropriately. The OpenSearch Software Foundation’s role is to provide exactly this infrastructure—a vendor-neutral home where enterprises can build, influence the project roadmap, and create their own ecosystems without proprietary constraints.

“Our role is to provide the infrastructure that companies can use to enjoy OpenSearch in an open way,” Lewis says. “They’re not on an island by themselves, because enterprises building on an open platform still want to know that they are part of a community that can provide the right support, the right versions, and the right security.”

Conclusion

As 2026 unfolds, the winners will be organizations that recognize agent-ready databases as the new baseline, maintain their foundation in proven search technologies, approach AI strategically rather than experimentally, and avoid vendor lock-in through open platforms. The OpenSearch Software Foundation is positioning itself as the community-driven infrastructure layer that enables all of this—providing enterprise-grade capabilities with the flexibility to scale alongside unprecedented innovation.