Welcoming our Computer Overlords
Are we on the cusp of intelligent systems autonomously running complex value chains?
Do they have the ability to apply logic and reasoning to isolate and investigate outliers in operating conditions and apply corrective or value enhancive actions?
Do we need to fear the imminent arrival of an army of planning and scheduling cyborgs from the future, intent on destroying every operational LP model that exists in spreadsheet format?
These are a few of the questions (save perhaps the last one) I have been thinking about recently. Mostly in the context of how will energy companies evolve to driving more efficiency and better returns from operations in a more dynamic and data-rich future. It has been an interesting discovery process as there does not seem to be a consensus view on this being a potential future state. However as one researches the elements necessary for such a future to occur, one can see the necessary prerequisites already in place. What remains, in my opinion, is a willingness to step into the slipstream of other industries that have already significantly evolved towards this future.
There is a common saying in trading “you live on the right-hand side of the screen”. If one visualizes a trader's price chart plot, the saying translates to the present being the very last column of pixels on the trader's computer screen. The future is opaque for it lies beyond the edge of the screen. The past, however, is fully on display in bright technicolor to the left side of the screen. The common challenge in trading is deciding how much does the past gives one insight into the next data point? And how does that insight, in turn, inform the next set of commercial decisions?
Historically, a fair bit of ink has been spilled on the topic of optimizing the returns from an asset position. Actively managing returns, commonly referred to as “Commercial Optimization”, is a set of processes that seeks to increase the return on capital by optimizing the economic and engineering efficiency of an asset through a planning and optimization cycle. In practice, this is typically measured by how the asset achieved or beat the planned (or forecasted) economic value expected over some defined future window as a result of re-optimizing the original plan due to changes in the market and operational dynamics.
This is the primer of a series of posts focused on how the confluence of distinct secular trends (relating to technology, commodity price cycle, human capital needs) and the business need to aggressively “sweat the assets” is going to drive the adoption automating tactical decision making for Supply and Trading organizations. We define these organizations as, those predominantly trading, converting, scheduling and transporting physical commodities in and out of a system of connected assets. The technology aspect of this inquiry is further made intriguing by the fact that everywhere around us it is increasingly deployed to listen, learn and react rather than record and process.
“The more information that’s out there, the greater the returns for just being willing to sit down and apply [analytics].
Information isn’t what’s scarce; it’s the willingness to do something with it”
Tyler Cowen, Average is Over
I will break this topic down into four dimensions in a series of posts.
The first dimension is to define the optimization problem across a sampling of energy value chains. This will give a view as to where autonomously driven optimization could work and the value drivers for adoption. After all, automation simply for the sake of it makes little sense unless there is a significant benefit to be gained.
The second dimension, which is a logical extension of the business case, defines technologies that will deliver truly competitive advantage and how to recognize them (Here are two hints: They already exist and they are not in the so-called ETRM/CTRM applications).
The third dimension is to look at the human aspect of this evolution. The driver and outcomes of automation in today’s and tomorrow’s labor force has some interesting tangents. More importantly, the required skills of the workforce best capable of driving value are already becoming evident. There will be a war for these resources when these skills gain industry recognition.
And finally, there is the "timing dimension", for lack of better terminology. Which is meant to articulate to the reader, why this is real and far from simply being an interesting thought experiment. Looking around, one can already see the advance scouting parties of the future amongst us and hear the distant rumble of destiny approaches.
While I am not predicting the rise of "self-aware synthetic intelligence systems" or recommend finding and investing in start-ups that are working on liquid metal poly-alloy compounds, I hope this series is thought-provoking and allows the reader to really examine the meaning of technological competitive advantage.