Quick
6 of 16 capabilities
Capabilities that compose over nothing, so no dependency expansion is needed.
Home / Enterprise Assessment
Enterprise AssessmentUnderstand what is governed, and what is not.
An assessment is the way into the platform. It reads the governance capabilities you already run, reports findings and limitations, and shows where every fact came from. It produces no score and no compliance verdict.
The way in
One assessment. Every capability that applies.
VerifAIer is deep infrastructure, and an assessment is how that depth arrives as one reading. It orchestrates the governance capabilities registered in your deployment, reads what they have already produced, by reference, never copied, and returns a single result. An executive summary and an engineer’s trace are two projections of that one result, which is why they cannot disagree.
Deployin your environment
Preflightscope, before anything runs
Assesscapabilities that apply
Readfindings and limitations
Traceprovenance to the source
Actin the capability that owns it
Reassesswhen you choose
Modes
Four modes. One pipeline.
A mode selects scope. It never changes what a capability reports, and a mode name is not a verdict.
Standard
13 of 16 capabilities
The default organization assessment, with declared dependencies expanded.
Deep
16 of 16 capabilities
The complete registered surface, including capabilities that cannot execute.
Continuous
13 of 16 capabilities
One explicit repeated-intent assessment using the Standard selection.
Applicable scope
Scope is measured, not assumed.
Nothing about your environment is guessed, and no capability is included because it sounds relevant.
Two observable questions
For every registered capability, discovery asks only what it can observe: does your context carry what this capability answers for, and does the capability’s own readiness check answer? Anything beyond that would be an opinion about a capability the assessment does not own.
Preflight runs nothing
Before anything executes, preflight reports what the request would cover: the mode, its selection, and any context a capability still needs. It calls no capability and produces no evidence.
Unreachable is reported, not hidden
A capability that cannot answer in your environment becomes a named limitation. It does not quietly drop out of the scope, and it does not fail the assessment. Everything else still runs.
Context-aware scope
A mode covers what applies, and says what it left out.
Read straight from the capability registry and the mode selection policy in the running product. Nothing here is a picture of the product; it is the product’s own answer.
Quick
6 of 16 capabilities
benchmarkcertificationcompliancedashboardflightgovernanceintegrationsmarketplacepackspassportreportsruntimesentinelsimulatorslazk
Standard
13 of 16 capabilities
benchmarkcertificationcompliancedashboardflightgovernanceintegrationsmarketplacepackspassportreportsruntimesentinelsimulatorslazk
Deep
16 of 16 capabilities
benchmarkcertificationcompliancedashboardflightgovernanceintegrationsmarketplacepackspassportreportsruntimesentinelsimulatorslazk
Continuous
13 of 16 capabilities
benchmarkcertificationcompliancedashboardflightgovernanceintegrationsmarketplacepackspassportreportsruntimesentinelsimulatorslazk
Solid: in scope for this mode.Dashed: deliberately out of scope, and named as a limitation in the result rather than hidden.
Reading a result
What the result actually says.
Complete is not full coverage
A COMPLETE result means every capability the mode selected succeeded. Quick selects six of sixteen, so its result always names the capabilities it did not cover.
Partial is a valid result
If one capability fails, the rest still run. A partial assessment is a working assessment reporting honestly, not a system failure.
Limitations are first-class
What could not be assessed is reported alongside what was: capabilities without a listing surface, unresolved dependencies, and scope the mode excluded.
Continuous is one run
The Continuous mode records repeated intent and performs one explicit assessment. It does not schedule anything and does not activate background monitoring.
Running one
How to run an assessment.
VerifAIer is local-first: the product runs inside your own environment, so an assessment runs there too, not on this website.
- Install or deploy VerifAIer in your environment: local, on-premise, air-gapped or sovereign.
- Sign in to the workspace with an account in your organization.
- Open Assessment at
/app/assessment, choose a mode, review the preflight, and run it.
What comes next
Every finding has somewhere to go.
A result is not a dead end. Findings carry provenance references back to the capability that produced them, so the next step happens in the capability that owns the fact, not in a report about it.
Evidence Engine
The deterministic record every capability derives from: envelope, receipt and provenance, verifiable offline.
ExploreCompliance
Evidence coverage mapped to EU AI Act, NIST AI RMF and ISO/IEC 42001 control packs.
ExploreTrust Intelligence
One agent-scope trust posture composing compliance, risk and quality.
ExploreAssess your AI operations.
Deploy VerifAIer and run your first assessment, or talk to us about enterprise rollout.