Komprise — Transparent File Tables (query-ready unstructured data to AI/lakehouses via Iceberg)

Tag: S-2026-06-23-komprise-transparent-file-tables Type: article (secondary news roundup + press-release coverage) Author(s): Tim King, Solutions Review (roundup) relaying Komprise press release; corroborated by BigDATAwire, Blocks & Files, StorageNewsletter, ChannelInsider, ITBrief. Date of source: 2026-06-23 (Komprise announcement; carried in Solutions Review roundup dated 2026-06-26) Date ingested: 2026-07-01 Authority weight: low — secondary/trade coverage of a vendor product release; capability and governance framing is vendor-relayed marketing, not independently verified. Raw file: S-2026-06-23-komprise-transparent-file-tables.md. External URLs in the raw stub.

What it claims

Komprise launched Transparent File Tables (early access, 23 Jun 2026), which indexes an organisation’s entire distributed file and object estate into a global metadatabase and exposes it to lakehouses (Databricks, Snowflake) as Apache Iceberg tables. Data teams can query Komprise-enriched, globally-classified metadata and pointer entries from their usual BI/analytics tools without copying petabytes of source files; when full files are needed for AI or analytics, Komprise Transparent Move Technology (TMT) with Intelligent AI Ingest moves only the required files (~2× the speed of standard transfer tools). Komprise states the offering provides “data governance based on user access permissions” — i.e. lakehouse access to the exposed unstructured data is tied to existing file-system entitlements. The stated problem is that unstructured data is ~80% of the enterprise footprint but <1% is used in AI (IDC, per Komprise).

For the vendor scan the relevance is that governed discovery and classification of unstructured estates, with entitlements preserved, is a control regulated FIs need to bring unstructured data into AI safely — bearing on GDPR sensitive-data discovery, BCBS 239 completeness/scope, and EU AI Act Article 10 data-governance for high-risk-AI training data. Komprise is an adjacent vendor (unstructured-data management / data mobility), not on the core catalogue/DQ/lineage/privacy watchlist.

Notable quotes

“…keeping governance tied to existing access permissions.” — Solutions Review roundup, 26 Jun 2026.

“Komprise provides data governance based on user access permissions.” — Komprise press release (via WebSearch snippet / trade coverage), 23 Jun 2026.

“Although unstructured data is over 80% of an enterprise’s data footprint, less than 1% of this data is being used in AI, according to IDC.” — Komprise press release (via WebSearch snippet).

What’s speculative vs. asserted

• Asserted (existence): Transparent File Tables exists in early access; exposes unstructured data as Iceberg tables to Snowflake/Databricks via enriched metadata + pointers; TMT/Intelligent AI Ingest moves only needed files.
• Vendor marketing (label as such): “data governance based on user access permissions”, the ~2× speed claim, and the IDC 80%/<1% framing — benefit/capability claims not independently verified.
• Not established / read-across only: how entitlements are enforced at query time; whether classification is defensible for GDPR/BCBS 239 evidence; no named regulated-FS reference customer; no mapping to a specific regulatory/evidentiary threshold.

Topics this feeds

• Komprise — primary vendor page (this release).
• Agentic Data Access Governance — adds an unstructured-data / storage-layer entrant to the agent-era governance category (governed exposure of the unstructured estate with entitlement-preserving access), distinct from the catalogue, access-enforcement, DQ and observability layers already mapped.
• BCBS 239 and Data Lineage — governed discovery/classification of the unstructured estate bears on BCBS 239 completeness/scope (read-across).

Open questions raised

• Does “governance based on user access permissions” enforce entitlements at lakehouse query time, or only at file-move time? The sources do not specify.
• Is the metadata classification auditable/defensible for GDPR sensitive-data or BCBS 239 completeness evidence, or only an operational discovery aid?
• Any EU data-residency option for the global metadatabase? Not addressed.