By Courage Or By Force - Machine Learning and AI Adoption In Oil And Gas

FACT: The rise of technology prediction will equal a drop in cost for gaining insight and access to information that drives strategic decision making within your company.

Over the past 5 years, my partner and I have invested our savings, kids college funds and paychecks into the development of something called Conversational Artificial Intelligence. Taking our combined 40+ years of field, office, sales, IT and accounting experience, we understand the impact this technology will have on this industry. This technology will place people in administrator role positions in a competitive crisis, as many of the tasks for gaining access, gathering intel and delivering results will become increasingly automated, significantly faster and cheap. Scary sometimes, but also extremely exciting to say the least. Coming from a loss averse industry that is blocking change like a 7 foot center in the NBA, part of us knows how much opportunity there is and we get as excited as a kid on Christmas at the sheer magnitude of efficiency gains and use cases that are possible.

Early in development we developed a product, found a field trial partner and launched a Conversational AI system with great results. The system was able to converse with field workers, create service tickets, dispatch workers, order products and services and answer questions within seconds, all while eliminating human interaction on the back side of the processes. We ultimately failed to get adoption at a rate that was sufficient to successfully deploy company wide and ultimately decided we were just a little too far ahead of the curve in order to launch industry wide at the time. This wasn't a one off client thinking about the use of Conversational AI in their business, we were actively speaking and engaging with several E&P companies and services providers at the time in order to move the product forward.

In fact, we were active in the oil and gas community speaking about the technology and applications, use and impact of Conversational AI technology within the industry.

Part of the problem with industry wide adoption is that there are natural trade-offs with the introduction of predictive modeling (AI) which is key to understand. More data means less privacy, more speed equals less accuracy, more autonomy means less control. Each company has to assess their key trade-off pros and cons in order to establish an idea of how and where to implement. Many will need to make decisions without full or complete understanding and information. This is a HUGE problem in the oil and gas industry. Most companies are largely loss-averse and will tend to remain inactive if they do not fully understand the technology. Remaining inactive may be their position, or safe-space, but failure to be active in adoption, could actually be worse than them being active in this brave new digital world.

Traditional loss-averse companies within the industry are now feeling the pain of making decisions by force, rather than by courage. As we continue to interact and engage with senior leaders across varying streams within the oil and gas ecosphere, we now see previously reluctant leaders turning towards technology adoption out of fear of being left behind, rather than courage to blaze a trail. They are beginning to understand that "something" has to change, but many still have no idea what this looks like, nor do they know what the pathway to this change looks like. To be honest, neither do we. We have an idea that successful deployment will be a cohesive and collaborative activity that will take both sides working together in order to reach the goal. Some of the barriers that once limited progress of adoption are now gone, and others are being removed every day.

Technological advancements have made things cheap that were once expensive. For example: Drilling engineers, managers, sales and support personnel attempting to extract critical data points in order to make mission critical decisions that can save or cost them hundreds of thousands of dollars still rely on data entry by real people, then time taken to aggregate, organize and send the data through email in order to receive the data required for making informed decisions. While this process has been made more efficient through the introduction of the internet and email, it's still far less efficient than that same engineer asking his Smart Speaker a question and receiving an answer back in seconds. With the addition of satellite and cellular access to job sites, sensors are now capable of transmitting large amounts of data to the cloud real-time. Connecting real-time data feeds to predictive models will give people access to insights they never knew, and with zero latency.

Data input has basically remained the same over the years with little change. With the ability to now connect multiple devices together and therefore connect Wellsite to the drilling project ecosystem, data entry can make a step change leap from where it is today. Cheaper predictive models will lead to more use of predictions. More use of predictions will require faster and more autonomous data entry into data lakes and stores in the cloud. This requires a new method for reporting entries that will not rely on people, rather on other machines.

At low levels, predictive machine models can relieve companies from relying on humans for tasks and therefore save on operational costs. As predictive models become more accurate and reliable companies will begin to utilize the models more and this will result in changing the way service companies do things. AI can effect the economics of a business so dramatically that it may change the way the company looks at strategy moving forward.

With all the struggle to maintain sufficient levels of profit for growth, expansion, investment and simply survival, companies can no longer afford to ignore the efficiency gains from predictive models. After spending the last 24 years working within this industry, I would bet they get there by force and not courage, but what do I know.

How will predictive models affect your oilfield business? Predictions influence behavior and decisions for all involved. The question is not whether or not we will get there? Rather, will we get there by courage or by force?