Introduction: Insights on AI adoption
Emerging technologies like GenAI and agentic AI have the potential to help public sector organizations increase efficiency through automation and strengthen decision-making, helping stretch tight human resources to better serve the public.
Procurement is the moment where AI’s potential becomes practical reality. This is the strategic process by which the organization identifies their needs, what they will buy, who they will partner with, and how the solution will be implemented to deliver maximum value for the public.
Done well, the procurement process can help the public sector secure the best solution to meet their mission as quickly as possible while mitigating risks, and strengthen productivity to better achieve their policy goals. Done poorly, the procured AI solution might be underused at best, or fail to achieve outcomes at worst, resulting in expensive failures that weaken trust in government.
To make the most of AI technology, everyone involved in the procurement process must be more nimble and collaborative than ever before. AI requires specific considerations to manage the unique risks and opportunities that come from the technology itself.
Leadership must step up to create AI governance policies and support agile procurement practices. Procurement officers need a basic understanding of how AI works, what they should pay special attention to, and whom to involve. Project and technology teams need to understand how the procurement process works, and what to consider as they collaborate together with procurement organizations. Everyone needs to be aware of what questions to ask, and when.
Insights on AI adoption in the public sector
During our research process, we identified several key insights into the current state of AI adoption that informed the design of this guidance, including:
- Hype vs reality: While building or customizing AI tech for cutting-edge use cases is getting a lot of hype, the public sector is currently mostly focused on buying off-the-shelf licenses for productivity solutions. Therefore, we highlight specific considerations for off-the-shelf AI purchases. Some vendors also make exaggerated claims around AI. We explain how having the right people on your team and asking smart questions can help mitigate this risk.
- Evolving attitudes towards risk: Most organizations are taking a very risk-adverse approach to AI, but early movers are now starting to roll back burdensome processes to enable easier access to low-risk technologies. We advocate that organizations take a strategic and risk-based approach to the procurement of AI technologies.
- AI adoption ≠ AI procurement: Procurement plays an important and under-appreciated role in AI adoption, but its role is limited with the trend towards centralized buying of off-the-shelf licenses and integration of AI into existing products. We provide recommendations to help organizations embed procurement in their broader AI adoption strategy.