Concentric AI
Type: company (data security posture management / sensitive-data discovery and classification) Sector: Data security, privacy and sensitive-data discovery First seen: 2026-08-28 Last updated: 2026-08-28
Created 2026-08-28 from S-2026-08-21-solutionsreview-dm-roundup (daily vendor-intelligence scan). Adjacent to the core watchlist — a sensitive-data-discovery vendor in the same category as BigID rather than a catalogue vendor. Single secondary source; all capability claims are the vendor’s own and unverified.
Snapshot
Concentric AI sells sensitive-data discovery and classification through its Semantic Intelligence platform. Its August 2026 signal is a vision-model capability that finds sensitive documents by their visual characteristics rather than their text — identity documents in particular — which addresses a discovery gap that text- and OCR-based scanning leaves open in image-heavy repositories [S-2026-08-21-solutionsreview-dm-roundup]. For EU/UK regulated firms this is a GDPR discovery-completeness question, since identity images accumulate in onboarding, KYC and claims stores.
Positions / Claims they advance
- Added a vision-model capability to the Semantic Intelligence platform that “identifies sensitive documents with consistent visual traits, including passport and driver-license images, even when the contents are blurred” [vendor claim relayed by trade press — S-2026-08-21-solutionsreview-dm-roundup].
- Positions the feature as “a lower-cost, more resilient alternative to OCR for discovering sensitive data in image-heavy environments” [vendor claim — S-2026-08-21-solutionsreview-dm-roundup]. No benchmark, dataset, accuracy figure or third-party evaluation is cited.
- Allows partners to build models for organisation-specific document types [vendor claim — S-2026-08-21-solutionsreview-dm-roundup].
Relationships
- competes-with → BigID — overlapping sensitive-data discovery, classification and DSPM positioning [inference].
- relates-to → Microsoft Purview — Purview’s own classification and DSPM stack covers the same discovery function inside the Microsoft estate; a firm would be choosing between or layering these [inference].
Tracked changes
- 2026-08-21 (reported week of; logged 2026-08-28) — Vision-model / visual-signature detection added to Semantic Intelligence, targeting identity documents including blurred images, with partner-extensible models for organisation-specific document types [S-2026-08-21-solutionsreview-dm-roundup].
Open Questions
- What are the accuracy and false-positive characteristics on identity documents? A visual-signature classifier that over-flags creates its own GDPR problem by surfacing and processing more personal data than necessary [inference].
- Is the claimed cost and resilience advantage over OCR supported by any published benchmark? Nothing cited.
- Where do the vision models run, and are document images transmitted off-premises for inference? Directly material to GDPR Art. 32 and to whether the tool can be used on EEA/UK identity data at all — not addressed.
- Does the capability produce exportable, audit-grade discovery evidence usable for GDPR Art. 30 records of processing, or only in-console findings?
- Any named EU/UK regulated-FS customer? None surfaced.
- Is Concentric AI worth promoting from adjacent to tracked-vendor status, or is this a one-off signal? Revisit if a second independent source appears.