Buying AI: Tips and tools for public procurement

The problem

Public servants are under pressure to make decisions about buying AI before the problems, risks and responsibilities are always clear.

• Procurement teams want help understanding, in plain language, what to pay attention to when buying AI technology.
• Procurement, project, and IT/data teams seek to better collaborate and work in more agile ways to get the best products and outcomes.
• Public sector organisations must ready their IT, data, and project teams to adopt AI technology.

The solution

Open Contracting Partnership developed a practical, plain-language guide for public sector procurement, project and technology teams. It builds on insights from more than 50 public-sector practitioners and experts and is designed to give teams a shared language and practical starting point, wherever they are in their AI procurement journey.

The guide covers readiness, common myths, buying pathways, solicitation design, supplier evaluation, contract clauses and contract management. It helps teams slow down and ask better questions before committing public money to fast-changing technology. This is designed as an ongoing practical resource.

The impact

The guide has been accessed by more than 30,000 visitors in the first month. A total of 500 procurement officials have been trained.

Lessons learned

Buying AI well is not only a technical task. It requires clear problem definition, cross-functional collaboration, risk assessment, supplier evaluation, good contract clauses and ongoing contract management. The rapid evolution of AI technology requires an agile approach to procurement.

This is a global guidance resource rather than a jurisdiction-specific project. It is designed to support public bodies operating under their own procurement, data protection, transparency, accountability and AI governance rules.