4C Associates’ Joe Temple considers the applications of AI in procurement.
How much does it cost to store a gigabyte of data? In 1980, the answer would have been hundreds of thousands of dollars. In 2017, the figure is closer to one cent. Improvements in computing efficiency have made it much cheaper for businesses to store and process data. Thanks to the rise of mobile cloud computing, expensive on-premise infrastructure is also no longer necessary. Businesses can now easily integrate data, leveraging the unlimited storage and intensive processing power of the cloud.
Procurement functions have been capitalising on these developments and are steadily adopting cloud-based, mobile enabled Source-to-Pay software. In their Q2 2017 report, Forrester estimated the eProcurement market to be growing at 11% per year. Process efficiency is an incentive, but equally as important is the ability to aggregate procurement data such as vendor details, contracts and purchase orders. This has laid the foundations for cognitive procurement: the application of AI to mine large volumes of procurement data and generate actionable insights.
So, what benefits can AI deliver in procurement?
Predictive insights for buyers
At 4C, we are already using AI to analyse large sets of spend data. Algorithms can automatically classify spend using a standard taxonomy: we use associative rule learning systems to allow businesses to compare their spend to category benchmarks from a representative cohort of companies, based on a range of selection criteria including sector, turnover and headcount. Savings opportunities are identified through Bayesian inference.
In March of this year, SAP Ariba announced that they were developing a procurement bot, leveraging machine learning to provide buyers with recommendations. Machine learning is the process by which algorithms iteratively learn from data: the more information fed into the algorithm, the more accurate the insight. As more and more data is channelled through SAP Ariba’s software, AI will be able to offer predictive insights with increasing confidence.
The applications of machine learning are compelling. Based on purchase orders, a bot could predict opportunities for consolidating spend. Using the results of past eAuctions, it could advise on the optimal number of suppliers to invite. SAP Ariba are not the only Source-to-Pay software provider developing this capability. Only this week, Coupa announced that it was acquiring Deep Relevance, a start-up applying machine learning to spend data to predict fraud.
Where predictive intelligence becomes exciting is in combining insight from internal company data with external sources, such as unstructured information on the web. 4C and Supplier.ai are currently collaborating to develop a tool for sourcing: the objective is to deploy AI to intelligently understand the product, geographic and contractual aspects of your supply base and crawl the web to identify potential supplier matches for you.
Upstream and downstream value
Requirements are being matched in the same way by AI further downstream in the Source-to-Pay process. Supply chain finance is being transformed by algorithms that can assess the likelihood of a corporate buyer paying a supplier’s invoice, providing banks with the confidence to offer suppliers significantly reduced payment terms. Currently, more than 40% of invoices raised by SMEs are paid late.
AI has the potential to create value throughout the entire Source-to-Pay process, but many procurement functions are rightly sceptical of its practical applications. This is where 4C can help. We can assess your organisation’s Source-to-Pay maturity and in doing so, help you understand what AI solutions are already on the market for procurement, what your AI strategy should be in the coming years and how you can begin to engage with this technology to retain your competitive advantage.
Contact Joe Temple at 4C Associates if you’d like to find out more about the applications of AI in procurement.