Your role as a procurement officer
| Who | Procurement teams |
| What | The role of procurement in supporting AI adoption in the public sector |
Procurement officials often play a key role in supporting AI adoption. You help define the use case, your decisions shape how AI systems are evaluated and implemented and what safeguards and monitoring are put in place, and you help ensure oversight over vendors. You play a critical role in enabling teams to move quickly on promising opportunities, test solutions before major investments, scale successful implementations, and adjust direction based on learning.
By focusing on organizational needs rather than technology hype, leveraging existing procurement best practices, and enabling agile implementation, procurement officers can help ensure AI investments deliver real value for their organizations and the public they serve.
We recognize that many of you are doing your best despite challenging circumstances, and some of the practices we suggest in this guidance might feel wildly impossible given your context. Even so, we encourage you to review the material and think about what you can do, and check out our Hacks for better AI procurement section for ideas on how to get started.
You are also not alone. You do not need to know it all or decide it all. Your project team and technology team are there to work alongside you. Here we break down what you ideally have decision-making power around, and what you can influence:
-
The procurement method and procedure: You can translate business requirements into effective purchasing strategies and flexible procedures that support innovation, and help experimental projects achieve better outcomes through flexible, iterative, agile approaches and pilots.
-
Timeline: Establishing procurement timelines and milestones that encourage strong and diverse vendor participation.
-
Evaluation model and criteria: For AI, criteria often go beyond price to include factors such as training data, risk mitigation, energy use, and much more. While procurement can advise technology and project teams on criteria, procurement can have the most decision-making power over the evaluation model.
-
Contract terms and management: For complex and evolving technology infrastructure like AI, proactive contract management is critical for monitoring performance, providing oversight, enabling iterative learning and adaption, and ensuring the best outcomes. You can help ensure prompt payment and compliance with laws and policies, and help structure contracts to enable scaling successful solutions. IT or the buying department or agency are often responsible for monitoring performance, but sometimes this is done by procurement, too.
-
Engaging the market and the public: You can recommend or decide to deploy strategies, such as user research with beneficiaries/end users and vendor engagement through request for information. These approaches are key to making outcome-driven projects that engage responsible vendors and deliver benefits to the public.
-
Transparency: Sharing information and crucial documents such as requests for information, requests for proposal, awards, and contracts, with the public can not only help vendors prepare better responses but also build trust with the public. Depending on your legal context, you can decide to proactively share more information through your existing public procurement and open data portals, but you can also create project specific public websites for more complex public facing projects.
-
A coherent and cohesive organizational AI technology strategy. In collaboration with the technology team, you can help build alignment around what technologies are purchased to avoid technology and vendor fragmentation, as well as avoid duplicate purchases across different departments or agencies.
-
Articulating the use case and what success looks like: A strong use case grounded in real needs is critical for determining what type of solution your team should pursue. You can also help support return-on-investment (ROI) calculations, and define vendor performance metrics.
-
Who is consulted as part of the procurement process, and when. You can help pull in experts at the right moments.
-
Requirements: Your requirements should support the responsible use of AI, and encourage participation through balancing risk with what the market can realistically provide.
-
AI governance and risk assessments: Your organization should have policies on AI governance and risk assessments in place, around how AI is used, and how to address risks around transparency, privacy, and bias. You can help ensure that vendors have the capacity to support responsible AI use through structuring appropriate contract terms and managing risk through procurement vehicles.
-
Understanding the market: You can help evaluate vendor capabilities and market offerings, including their relative maturity, as well as discover new vendors and emerging solutions.
-
Vendor relationships: Strong vendor outreach and engagement throughout the procurement cycle is critical for performance. You can also anticipate and support contract modifications as needs evolve.
-
How the AI solution is adopted and goes to scale: You can facilitate contract expansion for successful solutions, enable cross-agency adoption where appropriate, and help manage ongoing vendor performance.
