Supercharge your AI project with people-centered procurement

Although the United States spends over $170 billion dollars a year on critical social safety net services, accessing these services can be ludicrously burdensome. Technology, including AI, has the potential to improve how people experience these services by easing access to information for beneficiaries, empowering caseworkers to make more accurate and better decisions, reducing paperwork for all, and more.

If governments want to successfully leverage AI technology, they will need to reimagine their approach to procurement to put people at the center. Procurement is where policy becomes practice. Yet too often, procurement is an overlooked function, rather than recognized as a powerful ally for better government services.

We heard about the importance of shifting to human-centered procurement over and over again from U.S. State Health and Human Service directors, civil society organizations, and vendors who came together at the Center for Public Sector AI’s flagship convening to explore how AI technology can improve public benefits and service delivery. We also heard that dealing with AI can be intimidating for leaders who are otherwise experts in their fields.

The good news is that the challenges of procurement reform and dealing with rapidly evolving technology like AI can be addressed through some low-tech strategies – starting with the big first step of acknowledging just how important procurement is for safety-net service delivery.

Here are four people-centered practices that can make a big difference in your AI procurement: 

• Focus on real needs. What is the core challenge for people that you are trying to address through your project? Clarity around user needs will help you design a solution that works. This approach also allows you to explore different solutions and pick the best one to reach your goal…and the best solution might not be AI. 
• Collaborate with your tech and procurement experts. Right away, procurement and technology experts should be collaborating with whoever is seeking to make the AI purchase, and their involvement should continue through the evaluation and implementation phases. You don’t need to be a technical expert, but you do need someone with you who is, and who understands what you are buying and whether it’s working after it’s up and running. Procurement experts can help you write results-driven RFPs and get the most from your purchase. 
• Involve people along the way. AI projects will affect people responsible for managing the technology and using it, and especially those whose lives are directly impacted by AI-informed decisions. Consider how you might include all of these stakeholders in your decision-making process. This means finding opportunities for them to meaningfully participate, not just be informing them of what is going on.
• Know your data. A strong understanding of your data will help you set useful KPIs to gauge your project’s progress against your goals. Knowing your data will also help you proactively address the ethical considerations of your AI project. 

These people-centered strategies are the foundation for successful AI technology procurements. For more AI resources, check out our gentle introduction to applying AI in procurement, and stay tuned for fresh insights from OCP to come. To dive into more technical, step-by-step guidance for buying AI that touches upon critical questions such as IP, costs, and market engagement, check out great materials like the General Services Administration’s Generative AI and Specialized Computing Infrastructure Acquisition Resource Guide, and PUBLIC’s Buying Generative AI in Government guide.