PranavNagrecha/AwesomeSalesforceSkills / skills/admin/ai-ethics-and-governance-requirements
ai-ethics-and-governance-requirements
VerifiedUse this skill when defining AI governance policies, designing human oversight workflows, documenting bias mitigation strategies, or meeting regulatory transparency requirements for Salesforce AI features. Trigger keywords: responsible AI, AI bias, AI audit trail, AI transparency, human-in-the-loop, AI risk inventory, AI disclosure. NOT for the MLOps pipeline or Policy-as-Code engine architecture — use architect/ai-governance-architecture. NOT for Trust Layer technical configuration — use agentforce/einstein-trust-layer.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 14.8 KB |
| references/examples.md | markdown | 5.5 KB |
| references/gotchas.md | markdown | 6.5 KB |
| references/llm-anti-patterns.md | markdown | 9.4 KB |
| references/well-architected.md | markdown | 5.4 KB |
| scripts/check_ai_ethics_and_governance_requirements.pyexecutable | script | 9.4 KB |
| templates/ai-ethics-and-governance-requirements-template.md | markdown | 5.2 KB |
templates/ai-ethics-and-governance-requirements-template.md
View on GitHub# AI Ethics and Governance Requirements — Work Template
Use this template when defining or reviewing AI governance requirements for a Salesforce org. Fill in each section before starting implementation work.
## Scope
**Skill:** `ai-ethics-and-governance-requirements`
**Request summary:** (describe the governance task — e.g., "produce a governance policy for Agentforce go-live", "run bias evaluation on Einstein Lead Scoring", "design human oversight for NBA recommendations")
---
## 1. AI Feature Inventory
List every AI feature, Agentforce agent, and third-party LLM integration in scope.
| Feature / Agent Name | Salesforce Product | Customer-Facing? | Environment |
|---|---|---|---|
| (e.g., Einstein Lead Scoring) | Einstein Prediction Builder | No — internal to sales reps | Production |
| | | | |
| | | | |
---
## 2. Risk Tier Classification
For each feature, classify the use case on a 3x3 likelihood-impact grid.
| Feature | Affected Population | Likelihood of Harm | Impact if Harm Occurs | Risk Tier | Mitigation Owner |
|---|---|---|---|---|---|
| | | Low / Medium / High | Low / Medium / High | Low / Medium / High | (name, role) |
| | | | | | |
---
## 3. Responsible AI Pillar Mapping
For each in-scope feature, document the control for each pillar. Mark N/A only where the pillar genuinely does not apply and explain why.
| Feature | Accuracy Control | Safety Control | Honesty Control | Empowerment Control | Sustainability Control |
|---|---|---|---|---|---|
| | (grounding source, eval cadence) | (Trust Layer / toxicity threshold / content filter) | (disclosure copy + placement) | (override mechanism + escalation path) | (resource tracking approach) |
| | | | | | |
**Gaps (pillars with no current control):**
- [ ] (list each gap and planned resolution)
---
## 4. Human Oversight Design
Complete this section for every use case classified as Medium or High risk.
**Use Case:** _______________
- Oversight mechanism: (Approval Process / Flow review gate / Manual SLA-monitored queue)
- Reviewer role: _______________
- Review SLA: _______________ (e.g., 4 business hours)
- Backlog escalation: (who is notified when SLA is breached and what is the fallback)
- Log fields captured: Record ID, Reviewer User ID, Review Timestamp, AI Recommendation Shown, Final Decision, Override Reason
- Volume capacity check: (expected recommendations per day vs. reviewer capacity per day)
---
## 5. Bias Mitigation Record
**Model name:** _______________
**Model type:** (Einstein Prediction Builder / Einstein Discovery / other)
**Training data date range:** _______________
| Protected Attribute | Proxy Fields Identified | Proxy Fields Removed/Bucketed | Bias Report Run Date | Disparate Impact Score | Threshold Passed? | Sign-off Owner |
|---|---|---|---|---|---|---|
| (e.g., Gender) | (e.g., Job Title) | Yes / No | | | Yes / No | |
| | | | | | | |
**Retraining governance gate:** (describe the process — who triggers re-evaluation, where sign-off is recorded)
---
## 6. Transparency Disclosure Copy
For each customer-facing AI surface, document the disclosure language approved by legal.
| Surface (feature / page / component) | Disclosure Text | Legal Approval Date | Approver Name |
|---|---|---|---|
| (e.g., Product Recommendation Card LWC) | "These suggestions are generated by AI and are for informational purposes only." | | |
| | | | |
---
## 7. Audit Trail Specification
| AI Feature | Log Object / Mechanism | Fields Logged | Retention Period | Access Control | Gap vs. Field History Tracking |
|---|---|---|---|---|---|
| | (Field History + AI_Decision_Log__c / platform event) | (RecordId, ModelVersion, Score, Features, Timestamp) | (e.g., 7 years) | (Admin + Compliance role only) | (model version not in FHT — filled by custom log) |
| | | | | | |
**Incident response runbook location:** _______________
---
## 8. Regulatory Scope
| AI Feature | User Geography | Regulation Triggered | Risk Classification | Legal Review Required | Conformity Assessment Status |
|---|---|---|---|---|---|
| | (EU / CA / US-national / etc.) | (EU AI Act / CCPA / HIPAA / FINRA / none) | (EU AI Act tier or N/A) | Yes / No | (N/A / In Progress / Complete) |
| | | | | | |
---
## 9. Governance Policy Sign-off
- [ ] AI feature inventory complete and reviewed
- [ ] Risk inventory completed and owners assigned
- [ ] Responsible AI pillar mapping complete with no unresolved gaps
- [ ] Human oversight gates designed and tested for all Medium/High risk uses
- [ ] Bias detection run and documented; retraining gate defined
- [ ] Disclosure copy approved by legal and present on all customer-facing surfaces
- [ ] Audit trail specification complete and retention meets regulatory minimum
- [ ] Regulatory scope mapped and legal sign-off obtained where required
- [ ] Governance policy document signed off by legal/compliance and executive sponsor
- [ ] Review cadence scheduled (quarterly minimum; after each model retrain)
**Policy document location:** _______________
**Executive sponsor:** _______________
**Next scheduled review date:** _______________
---
## Notes
(Record any deviations from the standard governance pattern, interim controls, known exceptions with expiry dates, or open items requiring escalation.)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |