Wall St and White Collar Automation: Parallels for Procurement
White collar automation isn’t new but it hasn’t appeared in many professions yet. Wall St was the first domain to see it’s impact. For many decades, Trading in investment banks was dominated by a culture of aggression and it’s still very prevalent today although to a lesser extent. Male machismo and extravagance were emblems of power and dominance that were applauded in the top trading firms. But change was imminent and set to happen faster than anybody could have predicted back in the early 2000’s. Investment banking was adaptable and fast-paced so when smarter ways of making more money became possible, the Darwinian industry quickly evolved to more profitable strategies. That change was subtle at first but then hit the Trading profession like a freight train when it was clear that automating trading strategies was more profitable and less costly to operate. White collar automation was about to visit New York and London before anywhere else in the world.
The initial change started to occur from the turn of the millennium. A gradually dawning realization was emerging that Options pricing and estimating implied volatility in future prices of stocks and commodities was more significant than most appreciated. This was a competitive edge in the early days and motivated richer models for making smarter investment decisions. Traders could reason about a wider array of signals that were implicitly available in the market but it relied on colleagues with scientific skills to extract that value. In a sea of data, these quantitative models distilled valuable information to guide smarter trading decisions. Wall St employee profiles began to change. Traders were reliant on ‘Quants’ or scientists sitting nearby to model scenarios and provide richer market signals. Whilst the traders still commanded the larger salaries and machismo was still in vogue, the times were changing before the crash in 2008. The change accelerated when Hedge Funds realized that quantitative models could drive automated trading strategies with lower risk profiles and higher returns. High-frequency trading took hold and the share of trading volumes operated by manual decisions has been in decline ever since.
Procurement is following the same path. The Trading Agent Competition for Supply Chain Management was a precursor to this when university research laboratories developed a competitive game environment where manufacturing companies competed in a market for laptops. The input prices for transport, parts and commodities followed stochastic (random) processes with some underlying but discernible trends if analysed mathematically. Similarly, demand for different product types was somewhat predictable but again required scientific models to make smart decisions. This contest allowed AI research labs from 2008-2012 to build bots that would procure, sell and operate on behalf of a business. It proved that AI was ready to embrace automated trading in supply chain management and achieve results that would surpass what human experts could reasonably achieve.
Just as is the case with Wall St, procurement is playing catch-up with academic breakthroughs that are ready to be embraced and utilized. The primary barrier to adoption is the lack of appreciation for the significance of these AI-based techniques for wrestling with larger data sets and making more nuanced and better decisions.
There is an increasing flood of data that is available so procurement is searching for methods to embrace the increasingly complex world. Procurement teams have not had the resources to hire PhD graduates to implement internal models as Walls St did, the adoption of such levels of sophistication can still be realized via a more socialized model whereby tech vendors recruit such experts and develop AI systems to learn from data sets. Keelvar has a team of AI experts that are building Sourcing Bots to realize Intelligent Sourcing Automation for customers at the leading edge of procurement. The number of firms that will enjoy such advanced techniques will be small initially, but it will likely become a flood as soon as it becomes widely apparent how much more effective and efficient this approach is.
See Intelligent Sourcing Automation for further details.