This release centers on Semantic Model / AI Agent query capabilities and data security: the Semantic Model now carries richer dataset relation, classification, and value-range metadata; the AI Query API and MCP/Agent query capabilities have been expanded (median aggregation, hour/minute-level time granularity, and more); we've also rolled out HashiCorp Vault-based encrypted key management and broadly strengthened CKE (Consumer Knowledge Enrichment) data processing and sync pipelines, alongside fixes to several data management and cross-planet exploration display issues.
Release Date: 2026.09.17
Modules Version:
Upgraded Modules
Semantic Model Expansion
Dataset metadata now includes richer semantic descriptors — such as row-granularity notes, Known Discrepancies, and Disambiguation Rules — along with complete join-field information for dataset relations and a new Classification Rules capability, helping both users and AI Agents better understand what a dataset means and how it can be used.
Metric Management
Added a Metric creation/editing wizard with support for defining and validating Derived-type metrics, backed by new APIs — giving data managers more flexibility to define cross-dataset analytical metrics.
AI Query API / Agent Query Enhancements
The AI Query/Exploration API now supports median aggregation, finer time granularity (down to hour/minute) for filter, sort, and group-by conditions, and querying metrics directly by Metric ID — enabling AI Agents to handle more complex analytical scenarios. We've also added API support for virtual variable fields on relation datasets, and datasets are now flagged for raw-record-query eligibility so AI Agents no longer hit unexpected 403 errors.
HashiCorp Vault Key Management Integration