In the recent years, Oil & Gas and other critical infrastructure have become highly vulnerable to cyber-attacks. Most of us in Oil & Gas and in the Operational Technology (OT) world have probably heard about the Stuxnet attack on Iranian nuclear facilities. That was an inflection point and cyberattacks on Industrial Control Systems (ICS) and SCADA have only gotten more sophisticated since then. While there are many interesting and unprecedented ways that these OT systems can get attacked, one of the attack methods have been with the use of 'Botnets'. A botnet, in simple terms, is several internet connected computers with malicious software to create what is called distributed denial-of-service attacks (DDoS). The 'bot' within these botnets is a type of malware which allows the attacker to control an affected computer. Basically, this is what I will term a 'villain bot'. Their main work is to infect a host, spread themselves into the network and carryout tasks their master have programmed them to do - steal information, cause damage etc.
Well, that was enough of bad news and villainy. Now to some good news - with the emergence of Digital and Artificial Intelligence (AI), we have what I am calling 'Hero Bots'. We are seeing a surge of such bots in recent times that can automate and perform a variety of tasks within a process. With AI, these are starting to also include built-in learning capabilities. These heroes can mimic human conversations, analyse the conversation and perform a task such as providing contextual information, completing an e-shopping payment, or reserving a dinner table. They come in multiple forms:
Chat Bots: These are bots that can have conversations and respond to humans on a chat window. More recently these bots are transforming the e-commerce space; what is popularly coming to be known as Conversational commerce.
Task Bots: This is another set of bots that are driving a new trend called Robotic Process Automation (RPA). These task bots complete routine tasks assigned by users within a business process. With the rise of Cognitive technologies such as natural language processing and machine learning getting incorporated, these bots are starting to get new cognitive powers and are able to perform higher-order tasks.
Informational Bots: These are really like news bots. They can push personalized notifications while continuing to learn your preferences, browsing patterns, and information you may be looking for.
The further good news based on some publicly available information is that approximately 36 percent of all web traffic in 2015 was generated by good bots. At least 18 percent of all web traffic in 2015 was attributed to 'bad' bots, or bots created specially to harm sites, steal data, or perform other malicious acts. So atleast for now it seems like the heroes are winning.
The early adopter for the good bots have been in areas such as Digital Assistants in the Banking sector, Chatbots in Digital marketing, Conversational bots in the Retail sector or Patient care, and task bots employed in corporate function back-office processes in Finance and HR. One way of looking at this change is that the enterprise application products and solutions used within these early adopter processes are consciously embedding these bots as an application feature. So similar to how digital technologies such as Cloud, Mobility, and Analytics are transforming application architectures, bots are starting to become an integral part of enterprise products and solution architectures.
IT/OT convergence continues to drive and bring in all the best practices of Enterprise IT into the OT world. So, the question is - having suffered from the villain bots for all these years, is it time for Automation OEMs and System Integrators to start leveraging the 'Hero Bots' within OT products and solutions. Here are a few ideas for consideration:
Helper bots for the field operator: Mobile Operator Rounds software enabling digitization of refinery field operator tasks can further be enhanced with a bot companion that can use natural language processing and learning techniques to create Inspection reports, shift reports or automate tasks within maintenance and inspection workflows.
Project engineering bots: Data Historians, HMI/SCADA and control engineering toolsets could provide a configurable bot framework within the product to enable automation of engineering tasks - process graphics building, tag creation, linking tags to graphic objects, creating the control software documentation and others.
Test bots for major project upgrades: Test automation has existed for long. But what is emerging today in the Enterprise IT world is using AI-powered bots for software testing. The concept is to give a bot several thousands of examples of defects and intended correct functionality based on previous product releases. The bot could then generate test cases and auto-run these tests after major product upgrades and for regression testing.
Decision support bots: A pattern-detecting and continually learning bot could learn over multiple shifts of Operator actions and resulting plant and equipment performance to model a good performing shift and operator. These bots could then be used to train new operators or could provide intelligent and contextual decision support to operators to improve plant performance.
Oil & Gas industry leaders are pushing for and starting to develop an OT that is Open, Secure, and Interoperable. A key tenet of this future OT vision includes leveraging emerging Digital technologies across the solution layers. It may not be too early to include AI-powered bots within this future OT framework that will drive Automation OEMs and System Integrators to develop these Hero bots for the benefit of OTs.