Trust Computation

Formula

$ \text{effective\_trust} = \text{t} \times \text{authority\_weight} \times \text{temporal\_decay} \times (1 + \text{cumulative\_penalty}) $

Components

Base confidence (t): The raw trust score assigned to the claim [0.0 - 1.0].

Authority weight: Derived from the claim's tier:

Tier	Weight	Description	Example
1	1.00	Primary source	SEC filings, official records
2	0.85	Authoritative secondary	Analyst reports, peer-reviewed
3	0.70	General secondary	News, industry reports
4	0.50	Internal estimate	CRM data, team assessments
5	0.30	AI inference	Model outputs, extrapolations

Temporal decay: 0.5^(age_days / half_life_days)

If no decay field is set, temporal decay is 1.0 (no decay).

Cumulative penalty: Sum of all pen values from the provenance chain. Penalties are negative, so (1 + penalty) reduces the score.

Decision Thresholds

Score	Decision	Meaning
>= 0.70	ACCEPT	High confidence, suitable for decisions
>= 0.40	LOW	Moderate confidence, use with caution
< 0.40	REJECT	Low confidence, not suitable for decisions

Worked Example: Woodgrove Bank

Sarah's Claims

Claim 1: "Revenue was $4.2B" — t=0.98, tier=1 (SEC filing)

effective = 0.98 × 1.00 × 1.0 × 1.0 = 0.98 → ACCEPT

Claim 2: "Cloud growth 15-18%" — t=0.85, tier=2 (Gartner)

effective = 0.85 × 0.85 × 1.0 × 1.0 = 0.7225 → ACCEPT

Claim 3: "Pipeline strong" — t=0.72, tier=4 (CRM estimate)

effective = 0.72 × 0.50 × 1.0 × 1.0 = 0.36 → REJECT

Copilot's AI Claims

Claim 4: "AI adoption 34%" — t=0.78, tier=2 (McKinsey survey)

effective = 0.78 × 0.85 × 1.0 × 1.0 = 0.663 → LOW

Claim 5: "H2 will accelerate to 25%" — t=0.63, tier=5 (AI inference)

effective = 0.63 × 0.30 × 1.0 × 1.0 = 0.189 → REJECT

After Agent Consumption (penalty = -0.03)

Claims surviving threshold 0.5 filter: Claims 1, 2 (above 0.5 effective trust)

After -0.03 penalty:

Claim 1: 0.95 × 1.00 × 1.0 × 0.97 = 0.9215 → ACCEPT
Claim 2: 0.82 × 0.85 × 1.0 × 0.97 = 0.6760 → LOW

Temporal Decay Example

Claim 1 with decay=90 (quarterly):

Day 0:   0.98 × 1.00 × 1.000 = 0.980 → ACCEPT
Day 45:  0.98 × 1.00 × 0.707 = 0.693 → LOW
Day 90:  0.98 × 1.00 × 0.500 = 0.490 → LOW
Day 180: 0.98 × 1.00 × 0.250 = 0.245 → REJECT
Day 365: 0.98 × 1.00 × 0.060 = 0.059 → REJECT