Automation in energy trading today
In the energy market, Artificial Intelligence (AI) and Robotic Process Automation (RPA) are expected to transform operating models, and are forecast to be two of the fastest growth areas over the medium term.
When it comes to energy trading, RPA and AI are already adding value today, with companies actively replacing repetitive, process-oriented tasks. More often than not, value generation equates to cost reduction, with top line growth and quality of experience coming second. The emergence of other technologies such as blockchain will no doubt drive a shift in the focus of value generation at such as time that the core operations—often the priority of RPA and AI today—will cease to exist as a cost reduction opportunity.
Implications for energy traders
These changes in the use of technology in the front-office will have implications in terms of the skills and capabilities traders will need in the future. In energy trading, the typical 24×7 shift trading model could be a fruitful hunting ground for new sources of value from intelligent automation, where traded activity has a very different focus to that in the front-office.
Similarly, increases in the adoption of the Internet of Things (IoT) also act as an enabler to increased levels of AI across midstream activities, opening up new value opportunities across the enterprise.
Successful intelligent automation
Most organizations are grasping the basics through understanding the potential sources of value from automation, asking the following about their processes:
- Is the process intensively manual?
- Is the process repetitive in nature?
- Is there a high volume of work associated with the task?
- Is there significant probability of human error associated with the process?
- Is the process simple and does it have few exceptions?
However, the rise of the POC is rightly resulting in as much “fail fast” as “scale fast” as technology choices and capabilities prove problematic. In some instances, we’ve seen wholesale strategic vendor switches following prolonged issues attempting to scale certain technologies. In all cases, there remains a significant lack of skill available to implement rapidly or at scale and support these solutions effectively.
One of the challenges clients are likely to run into will be the fact that RPA and AI capabilities are frequently being looked at in isolation, rather than starting with the end in mind. The disciplines required to make RPA alone successful are not necessarily being adopted.
In an energy trading context, where significant value can be rapidly lost as a result of an inaccurate price, forecast or volume, adopting AI-driven automation on top of sub-optimal processes is very high risk. Changing behaviors and the workforce DNA thus becomes a crucial factor in success, where a business-led transformation will always yield greater benefits than an IT-led one.
Combining AI and RPA
While AI is at a much earlier stage of adoption today, it’s through its combination with RPA that clients start to yield breakthrough results. Today, even energy traders with strong analytics capabilities heavily depend upon human intervention to drive comprehension and action. Combining AI with RPA enables the pursuit of new value, with AI as the key enabler to a new, augmented enterprise.
Ultimately, AI consists of multiple technologies that enable computers to sense, comprehend, act and learn in terms of:
Energy trading transformed
Successful application of AI capabilities requires a high degree of technical proficiency and industry knowledge in order to operate and manage an algorithmic trading strategy. Having the personnel with the right trading experience to create algorithmic trading strategies and risk policy, the ability to manage and process much larger data sets used for predictive analysis, and being able to handle software configuration and hardware deployment are some examples of what’s required to operate and manage an AI-driven solution.
Thorough testing and rigorous management of those trading models also rise to the fore as crucial disciplines needed to manage greater complexity and reduce operational and financial risks associated with algorithmic trading activities.
Because most algorithmic trading models have a short lifespan, being able to do this continuously becomes key, where the ability of AI to learn and adapt is highly relevant—as is the importance of being able to rely upon robust process execution and integration across the trading value chain to provide timely and accurate data.
As organizations battle to overcome some of their legacy infrastructure challenges around the deployment scaling of a combination of automation technologies, it’s clear that it is only a matter of time before those challenges fall away and leave energy trading landscapes open to rapid improvement.
Chances are that as they do so, core problems with underlying data quality and timeliness will take over as the main constraints in the pursuit of value.