Series 3 — Part 3 of 6

Deal Intelligence Platform scores counterparty trust in B2B deals. This article covers the six weighted inputs, the trust deficit calculation, deal friction integration, and why "guarded evasive" is the safest safe-language label for a troubled counterparty score.

The Six Weighted Inputs

const TRUST_WEIGHTS = {
  communicationConsistency: 0.25,  // Does stated intent match behavioral patterns?
  commitmentFollowThrough:  0.20,  // Are stated commitments kept?
  informationDisclosure:    0.20,  // Is information shared proactively and accurately?
  negotiationTactics:       0.15,  // Are fair-dealing norms being followed?
  decisionAuthority:        0.10,  // Is the named decision-maker actually deciding?
  timelineReliability:      0.10,  // Are stated timelines accurate?
};

function computeCounterpartyTrust(engineOutputs) {
  const rawScore = Object.entries(TRUST_WEIGHTS)
    .reduce((sum, [signal, weight]) =>
      sum + weight * (engineOutputs.signals?.[signal] ?? 0.5), 0);

  const trustDeficit = computeTrustDeficit(engineOutputs);
  return {
    trustScore:   rawScore,
    trustDeficit,
    label: rawScore < 0.45 ? "guarded-engagement-posture" : "consistent-engagement-signals",
    requiresHumanReview: rawScore < 0.35,
  };
}

The Trust Deficit

The trust deficit is the gap between stated trust (what the counterparty claims) and behavioral trust (what they demonstrate). A counterparty who says "we are fully committed" while showing low commitment follow-through and high evasion patterns has a high trust deficit — and that deficit is often more predictive than the raw trust score.

Safe Language for Deal Risk

The governance wrapper converts all internal labels before they reach the Deal Intelligence UI:

  • deceptive → guarded communication posture observed
  • evasive → incomplete disclosure indicators present
  • untrustworthy → consistency variance between stated and observed behavior

These are not euphemisms. They are precise behavioral descriptions that carry no defamatory implication. See S3-ADV2 for the legal analysis.