# Public Agencies Set AI Pilot Graduation Tests

An article on public-sector AI procurement argues that agencies should set graduation criteria before deciding whether to scale, extend, or stop a pilot. It recommends testing systems against representative failures, documenting system configurations, and assembling a portable evidence dossier. For procurement teams, the approach shifts evaluation beyond demonstration performance toward measurable service outcomes, operational resilience, and credible plans for data portability, migration, and supplier exit. The article describes procurement approaches and policy direction; it does not identify a specific solicitation, award, or binding requirement.

- Procurement teams can define pilot evaluation criteria in advance, including how service outcomes and representative failure scenarios will be assessed.
- Contractors should be prepared to document system configurations and provide portable evidence that supports an agency’s scale, extension, or termination decision.
- Companies can address data portability, migration, and supplier-exit arrangements early in proposals and delivery planning; these factors may affect an agency’s ability to transition away from a pilot supplier.

**Jurisdictions:** international
**Industries:** Information Technology
**Topics:** Artificial Intelligence
**Published:** October 07, 2026

### Government Entities
- Indonesian government ministries
- Ministry of Digital Development and Information (MDDI)
- Singapore Ministry of Finance
- Lagos State Government

### Sources
- [Government AI pilots need a graduation test before they scale ](https://govinsider.asia/intl-en/article/government-ai-pilots-need-a-graduation-test-before-they-scale) - GovInsider