Naga Release

Release date: 25th February 2026

Smarter Duplicate Detection

We know how easy it is for content to be duplicated across an organisation’s information landscape. Documents are copied, shared, re-saved, and slightly edited over time, often resulting in multiple identical or near-identical versions scattered across systems.

The new Group By lens in Manage helps users instantly surface duplicate and near-duplicate content across their indexed data. By grouping search results using configurable entity values, the platform makes this duplicated content visible instead of buried.

For example, grouping on a record hash quickly reveals exact binary duplicates, while grouping on other entities supports broader comparison use cases such as identifying document variants or repeated communications.

This transforms duplicate detection from a manual, record-by-record task into a structured, insight-led process. Teams can reduce noise, understand how information is reused, and focus their attention on content that is genuinely unique and meaningful.

---

Making Patterns Obvious With Cluster Visualisations

Clustering and classification is powerful tool, but interpreting clusters through static tables can make it difficult to understand how documents truly relate to one another.

The redesigned Cluster Insights lens, available within AI Studio, introduces interactive visualisations that show how closely records related to each other, turning abstract model output into something users can see and explore.

Global views help teams understand how clusters relate across the full dataset, revealing natural separations and overall data structure. Local views make it easy to inspect individual clusters, identify outliers, review documents in context, and refine training sets by including or excluding items.

Combined with dimension reduction in the clustering pipeline, these visual tools support tighter, more meaningful groups and give users finer grained control over clustering outcomes. Instead of guessing why items are grouped together, teams can now see the relationships directly, leading to deeper understanding, better tuning, and more valuable outcomes.