Move over Alexa. Meet Winston. The artificially intelligent, voice-activated star of Dan Brown’s latest novel, Origin, can not only turn on lights and order stuff (including a fully crewed private jet), but also scrutinize vast quantities of data in seconds to determine the best course of action – such as which escape method Professor Robert Langdon should use to save his skin this time.
Unlike Brown’s fictional characters, sourcing organizations typically aren’t called on to make life-and-death decisions. Nevertheless, what procurement leader wouldn’t welcome having a Winston-like helper around to extract useful information, get things done and generally make their human team members smarter and more productive?
The CPOs of mobile network operator Vodafone and DSM, a big Dutch life and material sciences company, certainly do. During their keynote presentations at SCM World’s recent Live Europe conference in Barcelona, Ninian Wilson and Koen Devits, the respective CPOs at each firm, both referenced their use of Amelia, an AI-enabled “cognitive agent” that can answer questions about things like purchase orders or supplier invoices.
AI’s Impact on Procurement
Data from our latest Future of Supply Chain study shows that while three-quarters of practitioners expect transactional procurement processes to be largely automated during the next decade, less than 30% think that cognitive agents (aka chatbots) will add value, whether by answering enquiries from business users or supporting the work of buyers and category managers.
This percentage is likely to go up over time, as the use of such agents extends beyond the customer service center and into the back office, and their natural language capabilities improve. However, a glance at the other findings from this survey question reveal that AI in general is set to have a more pronounced impact on procurement functions (see chart).
Almost seven out of 10 participants – and 74% of those in procurement – think analytics and machine learning will routinely identify savings opportunities by crunching internal spend data and mining external market intelligence, whether in structured or unstructured form. Almost half (and 60% in procurement) see AI systems running elements of the sourcing process such as issuing RFPs and scoring supplier performance. And a third reckon such technology will even play a role in the development of purchasing strategies.
Indeed, while AI will undoubtedly lead to greater automation and fewer jobs in procurement organizations in the future (the subject of my previous blog), more significant in my view will be its ability to uncover insights that can be used to make better decisions and conduct sourcing activities differently – for example, by including a much broader range of potential suppliers in bidding events than is typically feasible today.
Just 34% of procurement professionals expect supplier rationalization to go into reverse in the next decade, as a result of cognitive agents’ contribution, our data shows. But this is actually quite a sizeable number when you consider that the technologies to enable such radically different processes and approaches – whether they be AI engines like IBM’s Watson or crowdsourcing platforms – are still being developed and are not widely adopted.
Equipped for the Job
Two vendor events I attended recently underlined the fact that while procurement digitalization is in its infancy, some companies are already experimenting with its tools. Cereal maker Kellogg’s, for example, has created automated one-year forward cost forecasts for key commodities using a mix of historical and external benchmarking data.
This type of analytics is “underleveraged in the procurement world,” noted Angel Mendez, a former Cisco CSCO and member of SCM World’s Executive Advisory Board. Speaking at the Cognitive Sourcing Summit in Palo Alto, California, he argued that a past lack of investment means that when it comes to the information at their disposal for negotiation purposes, “buyers are bringing a knife to a gunfight.”
Companies like Microsoft and Fitbit are seeking to redress this imbalance by using tools developed by Silicon Valley start-up LevaData, which hosted the summit. These analyse spend data, benchmark prices for electronic components and help buyers to determine when and where to run sourcing events. This digital augmentation is vital, said Rajesh Kalidindi, LevaData’s CEO, because “smart procurement people cannot cover all the bases,” given the scale, complexity and frequency of requirements today.
At the summit his company unveiled Leva, which it bills as “the world’s first AI advisor for strategic sourcing.” This claim aside, what’s interesting about this cognitive agent is that it seeks to go beyond process automation and administrative support, and provide buyers with useful intelligence and recommended actions in the categories and projects they work on.
As with any AI or any machine learning platform, these technologies feed on large quantities of high-quality data – something that most procurement organizations can only dream of today. Building a master data foundation is hard work, so any assistance that can help to get to the benefits faster, including from technology itself, is to be welcomed.
Like Amelia, Leva is not Winston. But it is a taste of what’s to come.