Supply Chain Talk: Optimising your supply chain with AI-powered spend analytics  

On 16 April 2025, SupplyChainTalk host Alastair Charatan was joined by Pam Wiseman, Principal - Supply Chain Strategy, Supply Chain Advisory Services LLC; and Greg Sloyer, Manufacturing Industry Principal, Snowflake. 

 

Views on news 

The head of the UK’s Competition and Markets Authority (CMA), Sarah Cardell, has warned of a "significant risk of bid-rigging" by contractors when bidding for public contracts. The CMA is testing  a new AI-powered tool to help detect such collusion. The trial programme, which uses AI to scrape large-scale data, is part of an attempt to reduce fraud and waste in the UK’s £300bn-a-year public procurement market. ML pattern recognition is capable of identifying collusion much better than humans.

 

High-value RFPs can be rather complex and full of costs and figures, which only an ML tool can monitor closely and consistently and combine with external data. An example for collusion in procurement can be when several companies agree which of them wins a certain part of the tender, while the others quote steep prices to make sure they let that particular company win. It’s more problematic for commercial companies to use the same AI tool, although once they’ve bought the tool and have the competence to use it, they can also apply it.  

 

Spend analytics 

AI has already been used for spend analytics for some time as a capability of procure-to-pay systems, which enabled procurement to break down spend by department, on-contract, off-contract, versus past and current budgets.

 

However, as always with analytics, data quality is key here. AI and LLMs today can help with contract analysis pointing out risks, as well as detecting discrepancies between purchase orders, planned orders, invoice data and receipts, where it can convert different units of measure and currencies in order to detect anomalies.  

 

Dashboards can present data differently than reports as they highlight the hot spots and elaborate on data that the user’s interested in in an interactive way. They also enable natural language interaction with data. When evaluating supplier performance, external weather data or global disruption alerts can explain why the supplier can not deliver consistently.  

 

The panel’s advice 

When monitoring suppliers, activate both leading and lagging indicators. 

Future solutions will be a combination of ML, GenAI on a solid data foundation.