Today we're announcing general availability of SWIRL Enterprise 4.0.
This release does something we've been working toward for a while: it renames our two core experiences to match what they actually do, and delivers meaningful upgrades to both.
Names that mean something
SWIRL has always had two distinct modes of interaction. One is search - fast, federated, ranked. The other is conversation - iterative, contextual, AI-driven. Starting with 4.0, they're called AI Search and AI Assistant.
The names sound simple. But naming things correctly matters. AI Search is for getting answers fast. AI Assistant is for working through a problem. They're different tools for different moments in the knowledge workflow, and now they look and behave accordingly - each with a clean two-tone layout designed around its specific use case.
The AI Assistant can write its own queries
The headline capability of 4.0 is that the AI Assistant can now generate queries in whatever dialect a source requires - SQL, SPARQL, MongoDB, Elasticsearch, OpenSearch, and any other query language the configured LLM understands.
In practice, this means a user can describe what they're looking for in plain English ("show me all Q3 purchase orders over $50K from APAC") and the Assistant will generate the correct SQL or structured query for the underlying source. The user never needs to know how to write the query. They just need to know what they want.
The same capability extends across human languages. Teams working in Japanese, German, Spanish, or any other language supported by the LLM can query in their native language - SWIRL handles translation at query time.
AI Search: confidence scoring and charts
AI Search now displays confidence scores alongside each result, so users immediately understand how strongly SWIRL ranks a match. For data-oriented SearchProviders, AI Search can generate charts from result sets - turning a search for sales figures or inventory data into a visual that's ready to share.
What this means for enterprise deployments
4.0 is the version of SWIRL that analysts, researchers, and knowledge workers will recognize as genuinely useful in their daily workflow - not just technically impressive. The gap between "what I want to know" and "what the system needs me to type" has closed significantly.
Version 4.0 does not require database migration. Full release notes and upgrade instructions are in the SWIRL documentation.