1997: IBM’s Deep Blue beats Garry Kasparov, the world champion, at chess
2011: IBM’s Watson beats legendary champions Brad Rutter and Ken Jennings at Jeopardy,
2016 and 2017: Google’s Alpha Go beats Lee Sedol, 18-time world champion, at the game of Go.
Besides beating humans at games, machines and computers are gaining cognitive abilities that make them better and better at more “serious games.” Artificial Intelligence (AI), after a period of stagnation, is entering a new expansionary phase. It represents a new wave of automation that is now blind to the color of the collar, be it white or blue. AI is making its way into the knowledge worker’s workplace, and the stakes are getting high.
AI is high on the agenda of many CPOs because it represents a new phase of automation possibilities and, at the same time, a new type of collaboration between humans and machines. Organizations can benefit from new efficiencies by further automating processes or through new processes to remove even more mundane tasks. In return, they can invest these gains to fuel effectiveness (delivery of value to the rest of the organization) by reallocating time to more value-add activities like managing relationships with stakeholders and suppliers. Also, they can enhance their value proposition by ramping up their capabilities in critical areas like insight development, decision making, risk management, and overall agility.
However, despite being a priority for many CPOs and Procurement organizations, the readiness is not there yet. Many teams still struggle with the “fundamentals” on top of which an AI can be relevant and bring tangible business value.
Data is more important than ever!
“In God we trust, all others bring data.” —William Edwards Deming, American statistician, professor, author, lecturer, and consultant, best known for the “Plan-Do-Check-Act” cycle
The case for data in business is not a new one, nor one that needs to be much discussed. The purpose is to collect sufficient data (and not too much) to make the right choices, with the maximum number of possibilities, while acknowledging that decisions are a sort of bet on the future and its uncertainties.
However, and as revealed by the progress at beating humans at various games, data is playing an even more critical role in the current phase of AI development:
- Deep Blue beat Kasparov for the sole reason that it could examine 50 billion moves in the three minutes allocated in a chess game. It was a super calculator built for a sole purpose. It won by “brute force.”
- To win at Jeopardy, Watson had to demonstrate a new capability: natural language processing. In a certain sense, it represents an evolution of Deep Blue as it combined computational power and “knowledge” with the need to understand the textual clues and matching them to answers in the game.
- Alpha Go is yet another step because the game of Go is so complicated. There are simply too many possible moves to be able to calculate the best one (the number of possibilities in the game of Go is often compared to the number of atoms in the universe but it is in fact much much larger). To win, Alpha Go had to learn and become a master itself. So, Alpha Go was not programmed to play Go but to learn it by playing it a lot and by learning from each game.
The above shows that technology is slowly but steadily having more human-like abilities. Instead of programming the machine with instructions to complete a task, the machine learns. Machine learning is a new programming paradigm, a new way of communicating your wishes to a computer that is deeply extending the scope and potential for automation and AI. This reinforces the role of data as basis to build a strong AI initiative. Several studies have shown that a poor learning algorithm trained on a high-quality and exhaustive dataset learns faster and produces better outputs than a better algorithm trained on poorer data! Attention to data quality is even more critical with AI / Machine Learning because the technology could amplify the underlying issues (biases or other) in the data it used to learn.
The lesson for Procurement organization is clear. A focus on data is a must and putting processes like S2C, P2P, and master data management in order is a prerequisite to leveraging cognitive applications. Digitizing the S2P process helps generate quality data and unify information (ex. a single supplier record spanning spend, contracts, POs and invoices). Master data management can help address existing issues in back end systems and unify information. Unfortunately, most procurement organizations still have a way to go to establish the solid data foundation required, as a recent study by Forrester shows.
Opportunities for Procurement
Organizations that are already on their “digital road” are in a good position to move to the AI era to address some critical business priorities and deliver a competitive advantage to the business.
There are two very important areas where AI and new digital solutions can support Procurement organizations:
- Enhancing “customer/supplier experiences”
- Making organizations “antifragile”
By liberating staff from more mundane tasks (thanks to, in part, better natural language processing and recognition, be it written, spoken, or via extraction in documents via intelligent OCR/data capture), AI can make jobs more human as collaborators will have more time to work with their extended network of partners, inside and outside. Also, technology will become more human-like and more adaptive/responsive to context because it will learn from each interaction. It will move from a role of “admin” to a role of colleague/consultant.
This role of assistant/consultant is, in fact, a matter of survival for companies. There are many business domains were the amount of real-time data to apprehend makes it impossible for us, people, to manage/decide anything. It is the case, for example, in risk management. Our world and modern supply chains are now far too volatile, uncertain, complex and ambiguous (VUCA). Therefore, teams must operate in a way that is smart, collaborative and agile. They can turn such capabilities into a massive competitive advantage by becoming antifragile and by striving in conditions that competitors cannot cope with. The same applies to many other decision-making processes where trade-offs between multiple criteria require deep analysis of several scenarios to be able to make informed and sustainable decisions.
Empowering, and enhancing procurement teams
Technology is a key enabler for organizations who want to strive in what Forrester calls the Age Of the Customer (AoC). Leading organizations are the ones who understand the place of technology: a means to an end, a way to enhance people, and a platform for growth. Leading organizations also understand what it takes to maximize the value of technology. They have, among other things, created strong foundations (process and data) that they leverage to move forward and prepare themselves for the future.