---
id: "framework-four-factors-trust"
type: "framework"
source_timestamps: ["§ 1. Measure Trust"]
tags: ["measurement", "metrics", "deloitte"]
related: ["claim-trust-roi-metrics", "action-measure-trust-factors", "entity-deloitte", "framework-five-approaches-ai-trust"]
steps: ["Humanity: Assessing whether the organization demonstrates empathy and fairness toward its employees.", "Transparency: Evaluating if the organization communicates openly and clearly about its decisions and tools.", "\\\"Capability: Measuring the organization's ability to deliver quality products", "experiences", "and functional tools.\\\"", "Reliability: Tracking whether the organization consistently keeps its promises over time."]
sources: ["adoption"]
sourceVaultSlug: "hbr-seg-adoption"
originDay: 9
articleStem: "hbr-edu-40-workers-dont-trust-ai"
sourceUrl: "https://hbr.org/2025/11/workers-dont-trust-ai-heres-how-companies-can-change-that"
sourceTitle: "Workers Don’t Trust AI. Here’s How Companies Can Change That."
---
# Deloitte's Four Factors of Trust

**Deloitte's Four Factors of Trust** is the behavioral and statistical framework at the core of the **TrustID Index** (see [[entity-deloitte-d9]]). It decomposes organizational trust into four measurable components, each scored on a **1-to-7 scale** based on employee agreement with specific statements. The composite is highly *predictive of behavior* — translating intangible feeling into actionable metrics for advocacy, engagement, and performance (quantified in [[claim-trust-roi-metrics]]). Deloitte has **open-sourced** the TrustID methodology.

**The four factors:**
1. **Humanity** — does the organization demonstrate empathy and fairness toward its employees?
2. **Transparency** — does it communicate openly and clearly about its decisions and tools?
3. **Capability** — can it deliver quality products, experiences, and functional tools?
4. **Reliability** — does it consistently keep its promises over time?

This framework is *Step 1* ("Measure Trust") of the broader [[framework-five-approaches-ai-trust]] and is operationalized by [[action-measure-trust-factors]].

**Expert caveat (from enrichment):** a single composite score can mask **domain- and context-specific** trust (e.g., trusting AI for spell-check but not clinical decisions) and is susceptible to social-desirability bias. Best practice pairs it with **behavioral data** (usage logs, error/near-miss reports) and **qualitative research** (interviews, ethnography) rather than treating it as a stand-alone KPI.