Static sensitivity labels do not survive contact with a document that keeps changing. AvePoint has released Kinetic Classification, which continuously re-evaluates data sensitivity across the content lifecycle rather than fixing a label at a single point in time.
The stated problem is discovery order. Security teams find out a file was sensitive after an AI agent has already read it. Kinetic Classification refines its assessment as new signals emerge, as access policies change, as agents interact with content and as labels expire.
The reach across applications matters because sensitive data no longer lives in one platform. Unifying sensitivity intelligence is intended to give security teams one picture rather than several partial ones.
The capability combines AI-driven and automated classification across both sensitivity and retention. It extends beyond Microsoft 365 and Google Workspace to GitHub, ServiceNow, Jira, Confluence, Box, Okta, Smartsheet, Monday.com, DocuSign, Bitbucket and AWS S3. The intent is one sensitivity picture across environments rather than separate answers from separate tools.
AvePoint chief product officer John Hodges framed the readiness gap. "Our 2026 State of AI Report makes clear that organizations aren't struggling with ambition; they're struggling with readiness: knowing what data is sensitive, who owns it, how it changes, and whether it should be used by AI in the first place," Hodges said.
That report found 82.7 per cent of organisations described themselves as very or extremely confident in preventing unauthorised data access. Seventy-two per cent of the very confident group still experienced an AI-related unauthorised access incident in the previous 12 months.
AvePoint chief technology officer John Peluso placed the release in the company's broader positioning. "Kinetic Classification and Rapid Recovery extend that foundation into the AI era, so organizations can classify, govern, and recover their AI estate with confidence," he said.
Peluso described the surrounding architecture as a connected layer of governance, security, recovery and backup controls sitting across an organisation's data. The company calls this the Trust Layer for AI.
AvePoint cites Gartner in support of the direction. As AI ecosystems expand across data, models and business processes, governance must move beyond static policies towards continuous, embedded and enforceable controls.
Recovery gets an order of operations
Announced alongside it, new intelligence in AvePoint's Rapid Recovery addresses what to restore first. Intelligent recommendations identify the most critical data to bring back. AvePoint describes this as restoring the minimum viable company when the starting point is unclear.
The Rapid Recovery Wizard lets teams pre-build and sequence recovery plans before an incident. Express Recovery for Entra ID adds the identity layer, building on Entra External ID protection announced in July.
The two capabilities are designed to work together. Kinetic Classification identifies where critical and sensitive data lives, so Rapid Recovery can prioritise what returns first.
The pairing is aimed at removing manual triage after an incident. AvePoint says organisations are instead equipped with a pre-built, sequenced recovery that a team can execute under pressure.
Both capabilities build on the agent governance and multicloud protection AvePoint added to its Confidence Platform earlier this year.