Trust Calibration Model

Trust Is the Primary Material of AXD

In traditional UX, the primary material is attention. In Agentic Experience Design, the primary material is trust. Every design decision either builds, maintains, or erodes trust. The Trust Calibration Model provides the vocabulary and measurement framework for understanding trust as a designable, measurable, dynamic property of the human-agent relationship.

Trust Has a Lifecycle

Trust is not static. It forms through initial interactions, strengthens through reliable performance, degrades through errors or opacity, and can be restored through transparent recovery. Each phase of the trust lifecycle has different design requirements. What builds trust in the formation phase is different from what maintains it in the maturity phase.

Calibration Prevents Both Over-Trust and Under-Trust

The goal is not maximum trust but calibrated trust - trust that accurately reflects the agent's actual capabilities and limitations. Over-trust leads to delegation of tasks the agent cannot handle well. Under-trust leads to unnecessary oversight that defeats the purpose of delegation. Calibration is the discipline of keeping trust aligned with reality.

Trust Signals Must Be Designed

Humans form trust judgments from signals - both intentional and unintentional. The Trust Calibration Model identifies the signals that influence trust formation and provides patterns for designing them deliberately. This includes competence signals (demonstrating capability), benevolence signals (showing alignment with user interests), and integrity signals (maintaining consistency and honesty).

Recovery Is More Important Than Prevention

Trust will be broken. Agents will make mistakes, misinterpret intent, or produce suboptimal outcomes. The quality of the recovery experience determines whether the relationship survives. The Trust Calibration Model treats trust recovery as a first-class design challenge, not an afterthought.

Provider Reliability Scoring

In agentic commerce, the agent must assess the trustworthiness of vendors, service providers, and counterparties on behalf of the human. The Trust Calibration Model provides frameworks for evaluating provider reliability based on delivery history, review patterns, dispute resolution quality, and consistency of service. The agent communicates its trust assessment to the human as part of purchase recommendations.

Brand Trust Translation

Consumers have existing trust relationships with brands. When an agent intermediates these relationships in agentic shopping, it must translate and preserve the consumer's brand trust signals. A consumer who trusts a particular retailer should see that trust reflected in the agent's behaviour - prioritising that retailer when quality and price are comparable.

Transaction Confidence Communication

For every purchase decision, the agent should communicate its confidence level - how certain it is that this purchase meets the consumer's intent. High confidence enables autonomous action. Low confidence triggers human review. The Trust Calibration Model defines how confidence is calculated, communicated, and calibrated against actual outcomes.

Post-Purchase Trust Reinforcement

After a successful purchase, the agent should reinforce trust through outcome reporting: what was bought, why it was chosen, how it compared to alternatives, and whether it met the specified success criteria. This post-purchase transparency builds the trust foundation for future delegations.