Classify and tag sensitive data, translate governance intent into enforceable policy, keep humans in the review loop, and constrain AI agents at runtime.
Govern enterprise data before it reaches AI, applications, or automated agents.
APIGQL is not only a federation layer. It connects sensitive-data discovery, field-level tagging,
governance policy, human review, drift detection, runtime enforcement, AI-agent controls,
and security risk intelligence into one continuous governance lifecycle.
APIGQL continuously checks whether architecture, human decisions, policy, and runtime behavior remain aligned. Architectural drift, governance drift, and XDR risk turn runtime compliance into measurable scores with explanations and remediation.
Architectural Drift
Reference Architecture vs. Current State
50%
Upload a reference architecture, design document, or approved target-state artifact.
APIGQL compares that baseline with the current architecture and highlights divergence.
Upload DesignExtract ControlsCompare Current StateScore DriftExplain
Example finding
Expected: API access routed through approved policy enforcement layer
Observed: one workload bypasses the expected control path
Drift score: 50%
Governance Drift
Approved Classification vs. Human Decisions
72%
Detect when reviewer actions move away from the approved governance baseline.
The score is accompanied by evidence and reasoning rather than treated as a black-box alert.
Example finding
Existing classification: Clear
Reviewer change: Sensitive
Reasoning: downstream masking and access policy will materially change
Governance drift: 72%
XDR Risk Intelligence
Security Logs → Risk → Remediation
84
Ingest security and runtime logs, correlate suspicious behavior, assign a risk score,
and produce evidence-backed remediation guidance.
Ingest LogsCorrelateScore RiskExplainRemediate
Example finding
Multiple failed privilege checks + abnormal data access pattern
XDR risk score: 84 / 100
Suggested action: isolate identity, review access token, tighten policy, inspect related traces
Runtime observability
Jaeger + OpenTelemetry make governance decisions traceable.
Policy decisions, agent actions, enforcement outcomes, and runtime events can be correlated with distributed traces,
giving teams evidence for why a request was allowed, masked, denied, escalated, or remediated.
This same execution loop powers GenAI, cybersecurity enforcement, compliance automation, and federated data access.
Select an industry to see governed AI in action.
MCP Orchestration
AI Agent Governance Engine
Specialized governance agents coordinate classification, retrieval, validation, and policy workflows, with MCP providing the interoperability layer for agent/tool orchestration.
Data Classification
Structured & Unstructured
Policy Synthesis
Governance → Runtime Rules
Snapshot-Grounded GenAI
Hallucination-Free AI
Output Validation
Policy & Fact Checks
MCP orchestration
MCP Context + Tool Orchestration
MCP standardizes how governance agents access approved context, tools, evidence, and runtime controls.
It works with the existing orchestrator rather than replacing it: the orchestrator coordinates agent execution,
while MCP provides a consistent interface for governed resources and tool calls.
1. Governed ContextExpose approved tags, policy metadata, evidence, and runtime state to agents through controlled context resources.
2. Governed ToolsClassification, policy lookup, validation, evidence retrieval, remediation, and enforcement are exposed as policy-controlled tools.
3. Agent OrchestrationThe existing orchestrator coordinates agent workflows while MCP provides standardized context and tool interoperability.
4. Audit + ObservabilityTool calls, policy decisions, reviewer actions, and enforcement outcomes are correlated with OpenTelemetry/Jaeger traces for end-to-end visibility.
Human-in-the-Loop + Explainable Governance
When controls trigger, teams can inspect decisions, ask “why,” and approve safe exceptions — without disabling enforcement.
Ask “Why?” Explanations
Policy • Evidence • Next Steps
Human-in-the-Loop Review
Approve • Escalate • Record
AI Agent Architecture + MCP Orchestration
Governance Agents • Orchestration