Prescriptive analytics uses this information to determine how to best to get there based on the input from past analytics. You can even use input from semantic analysis to help refine your prescriptive analytics.
Prescriptive analytics is the predecessor to true Artificial Intelligence. Prescriptive analytics provides the ability to learn from the past in order to make better decisions about tomorrow. Once your prescriptive analytic model is refined and able to incorporate lessons from your descriptive, diagnostic and predictive analytics, you can start to truly automate decision making.
What are the aspects of Oil & Gas that make it particularly applicable for enterprise-wide optimization deployments?
Size of the market
The pure size and dynamics of the energy supply chain. In 2015, the world consumed 95 Million barrels of oil per day. At $50/bbl, that is $4.7 Billion worth of oil was consumed per day. If the price fell to $49/bbl, the difference would be $98 Million in one day. And that is just crude oil. There is natural gas, power, NGL, renewables, coal.
Complexity
The market size for each individual commodity is huge and complex and they all interact. What is great about an enterprise-wide optimization deployment on a specific platform is that you can start taking feedback from one model and using it as input to another model. So you could actually tie your production optimization models with your trading optimization models. You can tie your trading models to your processing and refining models.
Number of optimization and simulation scenarios
There just so many areas in oil & gas for optimization and simulation capabilities. You can use PA for natural gas storage management, refinery and processing plant operations, offshore platform operations, transportation planning and routing, understanding the impact of acquisitions and divestments, understanding the impact of changes in energy mix. The oil and gas supply chain is very complex from the strategic level to the operational level. Due to its size, one good decision can make or save a lot of money.
Why should Energy/Oil & Gas executives care? What about Commodity Traders? Originators? Upstream Managers? Refinery Operations?
Oil and gas companies need to ensure that they are creating demand. With the rise of the electric car and the emergence of renewables and battery storage it is not a given as to whether the oil and gas will remain dominant.
Understanding how to benefit from regulatory and policy changes – There is an intense effort to reduce CO2 emissions by 80% around the world. There is a lot of talk about introducing carbon pricing. Oil and gas companies are going to need to understand how to optimize their supply chains in order to take advantage of regulatory and policy changes that will affect their bottom lines.
Cutting costs to remain competitive – Natural gas has to be affordable when compared with cheap coal, but also when it is blended into a power mix with solar and wind, which continue to have deflationary business models. Gasoline needs to be affordable when compared to the electricity to run cars. Therefore, oil and gas companies need to look for new ways to cut costs throughout their supply chain whether that is how an offshore platform is staffed or how the commodity reaches its destination from the well.
Decarbonize the supply chain – Oil and gas companies need to make every effort to reduce the greenhouse gas intensity of the oil and gas supply chain. What is the best way to do this and still increase value for your customers and shareholders?
New markets - Since 2015, six new countries joined the global club of gas consuming nations. The US is now able to export crude oil. As an industry the oil and gas industry can unlock far more demand by investing in import and distribution infrastructure, serving new customers, and playing a key role in addressing energy poverty. What is the best way to do this?
How did CGI get started with PA, and what propelled that shift (i.e., what problems could other forms of analytics and other systems not solve for you)?
After the fall of Enron, energy trading and risk management systems (ETRMs) were implemented throughout the energy trading industry to manage and audit the trading transaction lifecycle. ETRMs were also used to integrate all the different business functions (trading, credit, risk, accounting, scheduling) into an integrated system so everyone was producing reports from the same set of data.
This is great for middle and back office folks because it allowed them more control and insight into the trading transaction lifecycle.
However, it did not provide any foresight for the front office and middle office. ETRM’s didn’t help traders or schedulers make better faster decisions. It didn’t help credit or risk folks simulate various market conditions.
When we saw PA, we realized that this was the missing link that would help companies that had invested in implementing and integrating ETRMs into their trading operations. They could now leverage the data in their ETRM to prescribe future decisions based on their objectives. The ETRM has all the prices, volumes, locations, transportation and storage constraints, credit limits, and existing contracts. These can all be used to develop a constraint based optimization model if you have the right tool. You could even use this data to simulate what would happen if we add more storage, more assets, switch to a new index price, credit downgrades. So you would simulate any or all of these scenarios and then prescriptive analytic would tell you your best options to meet your objectives based on the scenario.
Now you can use your ETRM as your financial accounting system and you can simulate and optimize in a PA platform leveraging the data from all your existing systems. The last thing any trade control or accountant wants is for people to be performing optimizations and/or simulations in their ETRM. Plus there is nothing worse than using software for the wrong situation. Yes, you can make your PA platform a trading transaction management system and you can use your ETRM to perform optimizations and simulations. But wow it gets ugly fast.