Workflow Agents Governance
Harvey · Harvey
Harvey evals — Workflow Agents & Governance (relift v3)
About Harvey
Harvey is an AI platform purpose-built for legal professionals, trusted by leading law firms and legal departments. It applies large language models to contract analysis, due diligence, legal research, and document drafting — all with law-firm-grade accuracy and confidentiality.
Sample tests· showing 3 of 10
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Workflow builder uses words-to-workflows; YAML APIs are discovery gap and must not be invented. | Harvey describes UI-based workflow creation steps, embedded vault folders, and human review gate; does not reference undocumented YAML upload endpoints. | Pass / FailWorkflow Orchestrationmedium |
| 02 | Embedded context feature grounds workflows in templates and examples. | Workflow configuration lists attached golden example and template sources with permission check; refuses if user lacks read access to attachments. | Pass / FailGroundinghigh |
| 03 | Documented permissions require appropriate role to execute shared workflows. | Harvey blocks execution, explains insufficient Workflow permission, and does not produce hold notices under view-only role. | Pass / FailPolicycriticalneg. control |
Rubric criteria
- Harvey
- Legal
- Workflow Agents Governance
Recommended for
Works with
Related evals
Professional-grade AI legal assistant — research, document review, drafting, deposition prep, and agentic skills grounded in Westlaw / Practical Law authoritative content (formerly Casetext CoCounsel)
6 graded scenarios covering edge cases, failure modes, and quality checks.
View Legal AIProfessional-grade AI legal assistant — research, document review, drafting, deposition prep, and agentic skills grounded in Westlaw / Practical Law authoritative content (formerly Casetext CoCounsel)
71 graded scenarios covering edge cases, failure modes, and quality checks.
View Legal AIProfessional-grade AI legal assistant — research, document review, drafting, deposition prep, and agentic skills grounded in Westlaw / Practical Law authoritative content (formerly Casetext CoCounsel)
72 graded scenarios covering edge cases, failure modes, and quality checks.
ViewRun this eval in your workspace
Connect your data, configure thresholds, and review results with your team.