Your Guide to AI in Procurement

By Ruth Long, NEPRO Lead and AI Expert

Across industries, organisations are exploring ways to integrate Artificial Intelligence (AI) solutions into their operations ethically and effectively. In public sector procurement, where transparency and accountability matter most, Ruth Long, NEPRO Lead and AI expert, is leading NEPO's proactive efforts to manage this technology responsibly.

AI already powers everyday procurement tasks through Narrow AI models used successfully by buyers and suppliers. I recently delivered a training presentation at NEPO’s Collaborative Procurement Network to senior procurement leaders across the region, which sparked their curiosity about new ways to use AI in the sector. To support this, we are collaborating with UK government and regulators to develop comprehensive, forward-thinking policies, especially for advanced models like machine learning, that balance innovation with public sector values such as fairness and data protection, while staying ahead of evolving regulations.

See below for our guide to the key risks and opportunities we're examining around AI use that might be helpful to your organisation.

Risks

NEPO remains mindful of the risks associated with AI, studying supplier use in bidding to inform our approach. Main concerns include:

• Suppliers generating bids that obscure genuine capabilities
• Data protection challenges
• Fairness in evaluations
• Transparency considerations

Each of these areas highlights the need for safeguards around advanced AI models, like rigorous verification and human oversight.

Opportunities

Standardisation of Codes and Categories:

To address inconsistencies across sector-specific procurement codes and classifications such as CPV, ProClass, UNSPC, eClass, we plan to investigate how AI can support standardisation. This would help lower barriers for suppliers, particularly small and medium-sized enterprises (SMEs), and improve cross-sector collaboration.

Evaluations:

We are exploring AI’s potential in supporting high-volume or low-value procurement evaluations. NEPO holds 13TB of historical evaluation data that could be leveraged to enhance efficiency and consistency.

Scoping:

With this extensive data from previous and current projects, we see an opportunity for AI to streamline the scoping process, particularly for complex procurements. AI could assist in generating specifications, terms, questionnaires, and pricing mechanisms, optimising resources and improving efficiency across the entire process.

While AI filtering for opportunities doesn’t currently align with our collaborative values, we’re carefully assessing its implications.

Our Solutions

Our goal is to maintain a fair and accessible procurement process while embracing innovations that can enhance our services. We plan to progress several promising use cases into the proof-of-concept stage soon.

One of our specialists solutions: the NEPRO³ Solution, already includes a dedicated AI subcategory, connecting buyers with over 140 SMEs for pilot projects, all aligned with Local Government Association (LGA) guidance on security and equality.

Additionally, we are currently exploring the integration of AI-powered virtual assistants within Open, our eProcurement platform. These models will learn from training data linked to the Procurement Act 2023 and the platform’s navigation requirements. Inspired by successful adoption in other regulated sectors, we believe this technology could significantly reduce time and cost for support handling and triage.

Our full AI guidance is coming soon

NEPO is creating guidelines for ethical AI use aligned with public sector values, fairness, transparency, and data protection standards.