30 Days Gen AI Risk Trial -Start Now
Skip to main content
AI Security Coverage

AI security coverage: discovery, prompts, and files

Use this factual reference to separate a known AI tool from observed employee use, prompt controls, and file controls. Every decision still needs the exact application, action, endpoint, and deployment path.

For technical evaluators who need to map a proposed environment. This is a public, bounded reference—not a universal feature matrix or a substitute for a current test.

Evaluation workflow

Read coverage by state, not by a green tick

Start with the employee surface. Then separate four different facts: a tool may be catalogued, usage may be observed on a deployed route, a prompt action may be evaluated, and a file action may have documented scope. None automatically proves the next state.

1

Name the environment

Bring the app, employee interface, OS/browser, deployment option, action, and file type being considered.

2

Classify the state

Separate catalogued, observed, prompt-action, and file-action evidence instead of treating them as one coverage claim.

3

Close the gaps

Run a current synthetic test for unverified paths and record the result with an evidence date.

Coverage states used on this page

Current public scope · confirm in a synthetic evaluation
  • Catalogued: the AI tool is recognised in Aona’s 10,000+ tool catalogue. This is not usage observation or enforcement.
  • Observed: activity can be evaluated only on the agreed, deployed endpoint path; it is not a claim about every device or app interface.
  • Prompt/file action: assess the exact application, surface, data type, and policy response in a current synthetic test.
Bounded reference

Public coverage states

ScopePublic stateWhat to verify
AI tool catalogueCatalogued · 10,000+ toolsCatalogue recognition is not observed use or enforcement.
Employee activityConditional · managed endpoint pathBrowser, OS, deployment state, and the exact surface.
Prompt actionEvaluate per app and actionTyped/pasted path, policy response, and recorded evidence.
File redactionDocumented formats · DOCX, XLSX, PDFUpload route, modified-file behavior, and unsupported or unusual content.
macOS managed installNot shipped at last verificationCurrent MDM availability before adding macOS to a pilot.

This reference is deliberately bounded. Product and security should validate the exact row for a proposed environment and record the evidence date before a coverage claim is used in a rollout decision.

Important boundaries

No unexplained blanks should be read as support. An unverified path remains unverified until product and security review it.

DOCX, XLSX, and PDF are documented redaction formats; file behavior still needs route-specific validation.

The catalogue is not a universal-enforcement list, and the page does not claim app, OS, browser, plan, or native coverage that has not been verified.

Buying questions

Questions to resolve before a pilot

It means the tool is in the Aona catalogue. It does not mean the tool is observed, allowed, blocked, or otherwise controlled on your devices.
Qualified evaluation

Map coverage for your environment

Bring the app, surface, endpoint, policy action, and file type you need to evaluate. We will map the current evidence and name what remains unverified.

AI Security Coverage: Discovery, Prompts and Files | Aona