Since late last year, Royal Dutch Shell plc and Baker Hughes have been testing out Amelia, a cognitive knowledge worker developed by New York-based IPSoft. According to the firm, Amelia – a new artificial intelligence (AI) platform – could industrialize on a widespread basis functions around IT in the oil and gas industry that are primarily addressed through human labor.
Shell and Baker Hughes are using Amelia to deal with accounts payable in their customer service area. However, the potential for how Amelia can be applied is very broad, and she could indeed support operations out on oil rigs.
“For example, an engineer could interact with Amelia in order to speed up diagnosis and resolution of problems with machinery,” company spokesperson Sean McIlrath told Rigzone in an interview. “Understandably, however, even pioneering companies who are early adopters of cutting edge technology want to build familiarity with its capabilities in controlled environments first.”
At Baker Hughes, Amelia is being trained to interact with the companies suppliers in order to answering their invoicing queries, a time-consuming task due to Baker Hughes’ thousands of suppliers. By year-end the company hopes to be able to share data about how great an impact that Amelia has had on Baker Hughes’ operations, said McIlrath.
Unlike Watson – a supercomputer whose capacity is being tapped by Australia-based exploration and production firm Woodside Energy – Amelia is not meant to crunch data or perform analytics in a way that a human couldn’t do. Instead, Amelia is designed to do human work, and to learn through repetitive tasks, said McIlrath.
The impact of Amelia on human labor costs can be massive. Instead of involving human labor and using thousands of hours to close out tickets events, Amelia can handle tickets and events in a few minutes, said McIlrath. AI systems like Amelia are able to carry out many of the repetitive tasks that take up huge amounts of time and free up the workforce to apply its minds to more complex tasks.
“Importantly, we’ll leverage technology to change the way we want to run a process and invent new processes.”
AI won’t remove the need for human labor altogether, but will transform how human workers work, McIlrath noted. AI will allow accountants armed with spreadsheets and calculators to be productive in a way that they could not be when working with an abacus. “As machine technology evolves, the cost and design of the items we manufacture is transforming and the people involved in designing those products and running those plants are developing their skills all the time.”
While digital labor could transform a company’s operations, the inherent problem that has prevented companies from driving automation in digital labor is the lack of visibility across an enterprise and a lack of connection between differing systems. The Amelia platform effectively unifies all these systems, McIlrath noted.
“Everyday people use smartphones to gather information or make a transaction that was not even imagined 50 years ago – we’ll be using AI to redefine how we run operations in the same way. As we embrace and master this technology it will impact the skills we deploy across organizations,” McIlrath said. “The more we digitize the more we are able to analyze and act on the intelligence we extract from that data – this will permeate across the organization and affect all functions, not just IT.”
“The greatest benefit is being able to make sense out of all corporate data to make intelligent business decisions,” said McIlrath. “For example, in the case of data pulled from in-the-field sensors, intelligent machines/big data can help companies intake, analyze, process and output recommendations and insights that cannot be easily replicated by a human labor force because they must manually comb through all the data. Executives can get more accurate answers faster to make quick decisions that benefit the business.”
UNDERSTANDING HOW THE BRAIN WORKS
Developing the technology behind Amelia took time, said McIlrath. To create a technology capable of mimicking the human brain, the company’s research and development team spent a decade studying how human brains actually understand. Researching all aspects of how the human brain functions allowed IPSoft to define the right strategy for developing an artificial intelligence platform capable of understanding the meaning of what is communicated in natural language. This research also enabled IPSoft to learn how to follow a process and learn through observation following natural language, and independently being able to determine the steps needed to solve a problem.
“Once we developed the approach, we began writing the code,” McIlrath said, comparing the work to develop the technology to man’s ability to fly. “For hundreds of years, man could not fly because we were copying the movement of birds. Once we went deeper and actually understood the science of flight, however, we began building plans that could transport us over greater distances than ever before.”
Within IT operations there are highly defined processes, McIlrath said. The tasks that need to be fulfilled by IT engineers have been broken down into a virtual assembly line of operations so that the simplest and most repeatable tasks are performed by one group and as more complex tasks and exceptions are handled by other groups.
Within IT infrastructure management specifically, IPSoft’s autonomic technology, IPcenter, provides virtual engineers who can carry out L1 and L2 tasks so that more engineers are freed up to focus on L3 higher level tasks and even more importantly, resources can be refocused on planning and improvement of IT operations rather than “keeping the lights on” from day to day, said McIlrath.
“For CIOs [chief information officers], this is critical because they want to increase the budget they have for innovation and change which is only possible by reducing the budget on maintaining the legacy infrastructure running,” McIlrath commented.
As much as 80 percent of budgets are tied up in keeping infrastructure running. Focusing on digital transformation initiatives that can allow IT to drive significant business impact by enabling change in the cost, quality and speed with which actions can be taken.
“Essentially we are looking at shifting the balance of effort from a reactive modus operandi to a proactive, forward planning mode of working,” said McIlrath.
With an explosion in the amount of data that Internet of Things (IoT) brings into an organization there will be a need to measure, monitor and analyze that data – IT teams will want to leverage automation technology in order to bring that data together and big data analytics engines to derive intelligence from it.
On average, the company has seen mean time to response drop by more than 60 percent, while downtime and error rates on execution of processes drop to virtually zero.
“Ensuring higher availability of systems at all times naturally has an impact on the business. Also, with the proliferation of IoT devices monitoring equipment in the field a huge amount of new data needs to be transferred securely and combined with other systems in order to drive the level of intelligence required to operate a ‘digital oilfield’,” said McIlrath.
Ensuring all those devices are monitored and managed around the clock and that the data they transmit is made available for real-time decision making will be critical in order to remain competitive.
SMART MACHINES OFFER LONG-TERM COMPETITIVE ADVANTAGE TO OIL, GAS COMPANIES
Oil and gas companies urgently need to revisit cost structures due to the sustained decline in oil prices seen since June 2014. Smart machine technologies such as Amelia can offer oil and gas companies a chance to automate operations that will let them substantially increase their long-term competitiveness, according to a recent white paper by Enaxis Consulting.
According to Enaxis, the gains can be significant for those willing to challenge the traditional cycle of temporary cuts during oil price declines and then flooding resources into non-optimized operations.
“In fact, oil and gas companies can better utilize smart machines because of their structure and advanced data ingestion.”
Smart, connected technologies are now enabling a “third wave of IT-driven competition” that is likely to have a profound impact on most industry structures, according to a November 2014 article in the Harvard Business Review. Massive amounts of new machine-generated from embedded sensors, processors, software and connectivity or computers inside products, will enable many transformations in competition in most fields.
The first oil and gas companies to adopt smart machines will dramatically improve operational efficiency at significantly lower costs. Currently, the oil and gas industry is awash in data. To have a lasting competitive impact, much of the data needs to be cleansed and processed by either human labor or smart machine technology.
Two common reactions to the oil price declines that have previously occurred include mergers and acquisitions and the widespread adoption of Enterprise Resource Planning (ERP) systems.
“Through the use of machine learning and cognitive computing, new efficiency benchmarks can be set across a myriad of upstream processes,” Enaxis noted. “The result will be companies that are better able to mitigate future cycles of price fluctuations, increase resilience against market share attacks from their rivals, and enhance their competitiveness for the long term.”