A Cardiff data science firm has launched a whitepaper highlighting how artificial intelligence (AI) can enhance the integrity of public sector procurement in the UK.
The report – Risk Identification in Public Contracts: The Role of Generative AI and the Company Analysis Tool (CAT) – has been developed by Butterfly Data in collaboration with Oxford Insights.
It introduces the Company Analysis Tool (CAT) – a Generative AI solution utilising Large Language Models (LLM), designed to help procurement professionals detect red flags in awarded contracts.
With the UK Government ramping up efforts to address fraud and inefficiencies – particularly in the wake of COVID-related procurement scandals – the firm said the whitepaper “presents a timely and practical approach to leveraging AI for greater transparency and accountability”.
The need for AI in public procurement
Internationally, AI-driven tools like Brazil’s ALICE and the EU’s DATACROS II are already making strides in detecting financial misconduct. The UK, however, has yet to fully integrate AI into its procurement oversight mechanisms, Butterfly Data said.
André Petheram, Research Lead at Oxford Insights, has long believed that there is a huge opportunity to use AI to help procurement professionals monitor risky spending, saying.
“Procurement oversight in the UK struggles to keep pace with the complexity and volume of government contracts, leaving room for inefficiencies and possible corruption.
“We saw an opportunity to leverage AI to help procurement teams cut through vast amounts of data and highlight potential red flags quickly and effectively. That’s why we worked with Butterfly Data to develop the CAT, carefully designed to meet the real-world needs of public sector decision-makers.”
How the Company Analysis Tool (CAT) works
The CAT is designed to identify potential risks in public sector procurement by analysing company data from multiple sources, including Companies House, Contracts Finder, Wikipedia, and news outlets.
Explaining its functionality, Yash Yeola, Developer of the CAT and Consultant at Butterfly Data, said:
“The CAT is a powerful AI tool that streamlines the process of identifying high-risk companies in public procurement. By leveraging Large Language Model (LLM) technology, it synthesises insights from structured and unstructured data sources, flagging companies based on key risk indicators.
“These indicators include late or undisclosed contract awards, limited competition procurement procedures, newly registered or short-lived companies, and negative media coverage or financial irregularities. By ranking based on an entity’s risk profile, the CAT enables procurement teams to focus their due diligence efforts where they are most needed.”
Call for further collaboration
With an adaptable design, the CAT can be tailored to various government departments and regulatory bodies, enhancing their ability to detect fraud, ensure compliance, and protect taxpayer funds. Potential use cases for the likes of the Department for Work and Pensions (DWP), HMRC, and the Public Sector Fraud Authority are all outlined in the white paper.
As Butterfly Data and Oxford Insights continue refining the CAT, they are inviting government agencies, civil society organisations, and procurement professionals to contribute feedback and participate in further testing.
