Will Artificial Intelligence Impact the Procurement Function?


August 16, 2018

Interested in learning more about NPI’s services?

Contact Us

As AI (artificial intelligence) works its way into the enterprise, it gives rise to many questions. For many job functions, the question is “Will AI replace my job?” If your job is IT sourcing, that’s doubtful.

AI can potentially play a role in making the IT procurement process more efficient. It may have the power to expedite buying cycles, and give much-needed visibility into categories of spend that have historically been difficult to manage. It’s not likely that AI will replace the procurement function – especially IT procurement – altogether. Rather, AI could become an ally on the path to IT buying excellence.

It’s not about “Will AI automate procurement?” It’s about how to use AI to improve IT sourcing effectiveness.

AI relies on data. Good data. And it is our responsibility as IT sourcing professionals to help build the foundation for future AI implementations by working to gather that data. Today, those tasks include:

Contracts – An optimum system would have all contracts in a ‘readable’ format that can easily be indexed. As you’re working on agreements with vendors, request that copies of all documents are provided in a standardized format such as MS Word. Understand that many vendors place locks or other forms of security on their PDF files that prevent systems from indexing the content.

Quotes – Request that a spreadsheet file, such as MS Excel, accompany the proposal.  Ask the vendor to avoid any stylized formatting (skip the subtotals and extra lines of comments). You’re looking for a ‘simple’ spreadsheet that shows, line by line, each SKU you’re purchasing, the description, quantity, list price, and final price. Spreadsheet data will be even harder for AI systems to interpret unless these sheets remain simple and understandable.

Processes – In addition to good data, processes should be internally reviewed to help ensure that meaningful data is being collected in procure-to-pay systems. Is spend data capturing critical fields that future algorithms may employ? Examples include contractual start and end dates, termination notification cut-off dates, and cost center allocation data that will help to highlight specific areas where spend occurs (versus everything into a single ‘IT’ bucket).

Small steps with vendors and a conscious effort to build new processes that support future AI initiatives will help your organization move much more rapidly forward than your competitors when the AI day arrives.