Elevating data standards in the procurement function

Data quality and AI adoption

Modern procurement technologies are constantly enhancing their capabilities by adopting and leveraging advancements in AI. Applications of GenAI continue to be a big discussion topic for procurement leaders—and there are high-impact use cases to pilot next-generation abilities in the supply chain function.¹

However, according to our 2024 Deloitte global survey of 100 chief procurement officers (CPOs): “While 92% of CPOs are beginning to envision the possibilities of this new technology and plan to invest in it, only 37% were piloting or deploying Generative AI in procurement at the time of the survey". CPOs are still assessing the risk-reward tradeoffs as they begin deploying Generative AI.²

Surveyed leaders noted that data quality was one of the biggest obstacles to success and a major internal barrier to AI adoption in the procurement function.

Effective sourcing strategies, supplier relationship management, contract compliance, and risk assessment are just some examples of functions heavily dependent on accurate and consistent data. The need for accurate and complete data is particularly acute when implementing AI-based predictive modeling and scenario analysis.

AI-based intelligent analytics can provide meaningful inputs for key decisions only if the underlying data is accurate and comprehensive. Poor-quality data will likely lead to flawed recommendations and suboptimal decision-making in procurement processes with even the most advanced AI-based capabilities. For example, AI-based scenario modeling to identify risk in supply continuity depends on accurate supplier, inventory, and categorized spend data.

The topic of data quality isn’t new—but the time to build a foundation of best practices to prepare for the advancement of GenAI and help enhance end-to-end procurement operations is now. Below are two strategies to help you get started.

Start with effective content management

The procurement function operates on a variety of data and generates vast quantities of it—master data such as catalog items, units of measure and supplier information, and transactional data such as contracts, purchase orders and invoices.

Maintaining data quality is a persistent challenge that if not addressed can lead to missing, duplicate, or inconsistent data attributes. These data-quality issues are typically caused by limited data governance and/or reliance on manual inputs or disparate systems. And the impacts are substantial, from inaccurate reporting to ineffective decision-making.

Invest in data quality

This is one area where GenAI can be transformational, as deep learning models such as large language models (LLMs) hold the potential to significantly enhance data quality. We’ll outline some of the high-impact data enhancements for procurement that can benefit from GenAI. But first, let’s look at the key aspects of enhancing data quality and setting best practices.

• Data normalization: GenAI can automate the data cleansing process by identifying and correcting errors in datasets. LLMs can be trained on procurement-specific terminology and industry data to recognize and standardize inconsistent entries, and to identify and remove duplicative entries even when exact matches are not present.
• Data imputation: GenAI can address the problem of missing data and generate plausible substitutes for missing values based on existing patterns. Models such as generative adversarial networks and transformer-based architectures can be trained on existing data to learn complex patterns and correlations.
• Data augmentation: GenAI with models like variational autoencoders can augment procurement data by creating synthetic datasets to enrich data. This is also particularly useful for scenario analysis—for example, simulating the impact of a sudden supplier outage on costs and timelines, which allow an organization to develop contingency plans.³

Moving forward in the age of AI

Now is the time to invest in the foundation your enterprise needs to be ready to leverage the GenAI revolution. Procurement leaders can approach data quality through sophisticated AI tools for data cleansing, imputation, classification and augmentation. By addressing common data quality issues and providing more accurate, comprehensive insights, GenAI can empower procurement teams to make better decisions, reduce risks and drive efficiency. As adoption of AI technologies grows, we can expect a new era of data-driven procurement that delivers strategic value across the supply chain. Begin your journey by evaluating your current data issues and establishing robust data governance models, and assess solutions in the market leveraging AI-powered data management.

End Notes

¹ Ryan Flynn, Mike Deng, Vinay Rajani and Ayushman Kaul, “CPOs steering GenAI in procurement through uncharted waters,” Deloitte, August 21 2024.
² Ryan Flynn, Mike Deng, Vinay Rajani and Ayushman Kaul, 2024 Global CPO GenAI survey, Deloitte, 2024.
³ Spencer Young and Phill Domschke, “Planning your AI journey?,” Deloitte, June 11, 2024.

Get in touch