The GroundSync Score: Quantifying AI Trust
The GroundSync Score is a 0-100 metric that represents the percentage of AI-generated claims that were successfully verified against authenticated data sources. It's the first objective measure of AI response trustworthiness for personal context.
How the Score is Calculated
GroundSync Score = (verified_claims / total_claims) × 100 Example: - AI response contains 5 claims about personal context - 4 claims verified against calendar/email/financial data - 1 claim could not be verified (no matching data source) Score = (4/5) × 100 = 80
What the Score Tells You
Immediate Trust Signal
At a glance, you know how much of the AI's response is backed by your real data. A score of 95 means you can trust it. A score of 20 means proceed with caution.
Data Coverage Indicator
Low scores often indicate missing data sources. If schedule claims can't be verified, it may mean your calendar isn't connected yet.
Actionability Threshold
Organizations can set minimum score thresholds for AI agent actions. "Only proceed if GroundSync Score > 80" becomes a simple, enforceable policy.
Score Components
Each verified response includes not just the aggregate score, but a breakdown showing which claims contributed to the score:
{
"groundsync_score": 85,
"claims_breakdown": {
"verified": 17,
"suggested": 2,
"needs_input": 1
},
"by_source": {
"calendar": { "verified": 5, "total": 5 },
"email": { "verified": 8, "total": 9 },
"financial": { "verified": 4, "total": 6 }
}
}Use Cases
User-Facing Trust Indicator
Show users the GroundSync Score alongside AI responses so they know how much to trust it. "This response is 92% verified."
Agent Guardrails
Configure AI agents to require human approval when GroundSync Score falls below a threshold. Prevent autonomous actions based on unverified assumptions.
Quality Monitoring
Track average GroundSync Scores over time to measure and improve the quality of AI-assisted workflows in your organization.