Governance

Govern hundreds of agents without losing control

When every team builds its own agents, quality drifts. Pivotal Bench gives CIOs and AI teams a shared, measurable standard.

01

Craft.ecQ score from 1 to 5

Every agent is rated on 8 criteria: Context, Role, Actions, Format, Tone, Examples, Constraints, Quality control. Red, orange or green badge.

02

Criterion-by-criterion explanations

Present, partial or missing, with a short reason: agent builders know exactly what to improve.

03

Manual or automatic evaluation

In automatic mode, each agent is evaluated when opened and re-evaluated when its instructions change. Results are kept.

04

Multi-company organisation

Company recognised by e-mail domain, member approval, extra domains, Google or e-mail sign-in.

05

Roles and per-space sharing

Invite colleagues as readers or editors. Everyone only sees spaces they created or were invited to.

06

Controlled prompts and models

Admins pick the evaluation model and tune the evaluation and Craft.ecQ prompts in the company back office.

Your AI agents deserve serious measurement

Connect your Dust workspace and get a first inspection and a first benchmark within minutes.