With today’s almost endless supply of storefront and online vendors, the role of procurement professionals is more difficult and more important than ever. They take many variables into account—timeframe, price, shipping costs, reliability of vendor and more—to get the goods their organizations need at the right price and the right time. These pros gather this knowledge over years of experience.
This wealth of knowledge must be considered as organizations create procurement policies. Otherwise, blind enforcement of generic policies may result in people finding ways to go under the radar. Purchasers sitting within their silos make isolated purchasing decisions based on their experience alone rather than the collective experience of the enterprise. In addition, there are always some who try to game the system.
Though tighter controls may yield more efficiency, they’re not necessarily the solution because they invariably restrict business agility. To accelerate procurement without compromising efficiency, enterprises need to empower people to make quick purchasing decisions without losing control over how the money is spent.
AI Addresses Pain Points
The quality of intelligence available on purchase transactions is the main factor that makes procurement so difficult. It is dated by the time it is received, leaving little or no time for any kind of interception or guidance. Part of this intelligence is derived through traditional analytics, which employ a slice-and-dice approach to analyze data. They help understand the spend distribution over a period of time and identify opportunities for optimization.
But these analytics aren’t able to look deeper to find the patterns of buying behavior that may need to be probed, encouraged or stopped. Intelligence is also derived from subject matter experts or consultants who analyze the spend distribution and provide advice based on industry benchmarks and best practices. However, they fail to drill down to a transaction level and provide specific recommendations.
With the vast quantities of procurement data now being generated, it’s essentially impossible to analyze it all to truly understand what’s happening. Purchase transactions have patterns hidden deep within them, some of them good and some bad. These patterns reveal the nuances of buying behavior, and they constantly evolve. The problem is that you don’t know what they are upfront. Hence, you cannot define any rules to detect their occurrence. That is why traditional slice-and-dice approaches fail – and why AI is so helpful.
AI increases the transparency in procurement by enables staff to see what is not readily apparent. It can auto-discover patterns in purchase transactions that look odd using algorithms and then highlight them to humans. It can observe and learn which of those patterns are accepted by humans as worth monitoring through feedback loops. It can then use this knowledge to detect and predict anomalies in live transactions, allowing humans to intercept and take timely action. That’s when the procurement function starts to become cognitive.
Dealing with Exceptions
In the procurement department, exceptions come in a variety of guises. Some adversely impact spending because of avoidable price variance, some impact the cost of operations because of avoidable delays and some are non-compliant with procurement policies. Exceptions can be positive as well, such as transaction sets that are always compliant and never result in price variance or delays.
To get to the root of an exception, it’s necessary to first identify an outcome that defines the exception and then identify a set of influencing factors that could produce it. The outcome could be price variance, which is the difference between the price quoted in an invoice and a standard price at which the item might be bought. There could be any number of influencing factors behind such an exception: business unit, plant, buyer, supplier, item, time of the year and more.
You would be able to define such an outcome, along with any number of influencing factors, by incorporating AI into the procurement process. This can also help predict likely exceptions ahead of time. Sophisticated algorithms then take over to crunch a purchase transaction data set and discover patterns that require inspection and are presented to humans with transactional evidence. Such a virtual procurement expert would be able to compute and present a financial impact of every identified pattern. Then human procurement experts can validate these patterns.
Partnering humans with AI in the procurement department could then be used to discover what drives other types of exceptions, such as transaction fallouts, mavericks, anomalies or the unavailability of a purchase order against an invoice.
Benefits of AI in Procurement
Speed is required for rapid business growth, and AI empowers procurement professionals to operate at such speed and with greater efficiency. When a layer of intelligence is always at work, organizations can continuously monitor and guide people to make the right decisions based on the organization’s collective experience. Exploiting hidden opportunities to optimize spend by identifying and eliminating maverick transactions produces an efficiency boost.
Improved compliance is a happy side effect of AI. Instead of insisting people comply with a generic set of policies, the application of AI allows procurement teams to become more sensitive to real business needs. It enables them to continuously engage with people on the ground and help them make the best choices within their constraints while staying compliant.
Tucked away on the sidelines, procurement professionals work diligently to find all that an organization needs. They must source, price and order a myriad of goods, but if those processes take too much time, it hinders business growth. This is where AI becomes so helpful, eliminating mundane tasks and shifting the department mindset from focusing on process to focusing on data. The data provided by AI enables procurement pros to make more informed decisions that help the organization move faster and with greater responsiveness to market needs and business goals.
Akhilesh Tripathi is the global head of Digitate, a software venture of Tata Consultancy Services. He has led the management of strategic alliances with software vendors, and participated on the advisory councils of several strategic vendor partners.