As the Director of Analytics Solutions and Operations Manager in Houston, James Sokolowski advises on effective analytics strategies at Tessella
It is no secret that many industries are looking to developments in artificial intelligence (AI) and the internet of things (IoT) to reduce the cost and risk of identifying new opportunities. The oil and gas sector continues to drive down the cost of exploration, while pharmaceutical companies are battling with the fact that nine out of every 10 clinical drug candidates fail approval.
In many cases, the introduction of these emerging technologies is helping. The large legacy companies that dominate these traditional industries have been around for a long time and have a wealth of data at their fingertips, making it the perfect opportunity to take advantage of new technical developments to speed up and simplify processes. Pharmaceutical giant Sanofi has dedicated significant funds to the development of end-to-end analytics, whilst Merck announced plans to connect “smart factories.”
(Full disclosure: My company, Altran, worked with Sanofi on its innovation and collaboration strategy.)
However, some companies are finding the journey to digital transformation an uphill battle, or are left wondering why their efforts are not leading to any clear return on investment (ROI).
The Industry Myopia
One reason traditional industries are struggling to transition to AI and the IoT is that these companies have previously existed as isolated islands of expertise and have operated in their own “silos” of knowledge, only looking to the experts within their field to solve complex problems.
The reason these silos exist is that the sectors have developed separately from other industries, and thus, they have become reluctant to trust experts in other fields. Oil and gas companies previously assumed that the best people to design efficient and cost-effective methods of oil exploration are geologists or petroleum engineers. Pharmaceutical companies would have assumed that the best people to work out how to identify new drug molecules for treatments were analytical chemists.
As such, these legacy industries have evolved to accommodate only forms of expertise from within their own verticals, assuming that other sectors will not experience the same problems that they do. They developed a “vertical mindset” with preconceived ideas about how to solve issues rather than thinking bigger.
However, as industries such as health care and energy begin adopting digital technologies, they need to draw on expertise from other digital native sectors. The oil industry needs expert knowledge of oil exploration, but it does not necessarily have expert knowledge of machine learning or smart data. As the IT and software industries converge with legacy industries such as health care, firms need to be prepared to open themselves to outside expertise. An image-recognition algorithm designed for driverless car cameras might have applications for autonomously analyzing brain scans to diagnose disease.
Despite the stark differences among the manufacturing, oil and gas, and pharmaceutical industries, there are commonalities in the processes for these fields and how they draw insights from data to become more efficient. In a time when rapid adoption is imperative to success, businesses must look beyond superficial differences between sectors and instead focus on what they have in common.
Embracing Similarities Between Sectors
That is not to say that you should involve the experts any less -- you still need a petroleum engineer to develop drilling plans -- but multiple disciplines are vital to get the right answer quickly and efficiently. Only that way can the commonalities between fields and the broader connectedness of analytics be identified to help solve challenging problems. This means companies will not only need to adopt technology from other fields but also recruit a broader array of talent and bring in outside consultancy from other fields.
For instance, I have worked on projects where experts in anomaly detection techniques originally adapted for use in military applications were brought together to help aid autonomous monitoring of a nationwide fleet of gas turbines, in order to better track efficiency and also predict necessary maintenance. I have also seen experts in acoustic spectroscopy help pharmaceutical companies to discover data science methodologies to help better identify new potential drug targets more effectively. And data analytics techniques learned from the nuclear industry have been adapted to help maximize the yield from oil reservoirs.
Digitization, when done well, involves more than just reinventing the wheel for each individual process or application. Instead, the best data science comes from utilizing an existing bank of knowledge and expertise to go further with the processes that have been tried and tested in other industries, and what is already known to work. No less because coming up with something from scratch can cost you a lot more and take considerably longer than building on existing projects.
That is not to say that it is never suitable to start from the beginning -- this could help you to carve out something new and get ahead of competitors. It is fine if you have the time and money, and don’t need to worry about your competitors catching up. Using existing tools at your disposal is much more efficient and cost-effective.
In a highly competitive landscape, time is not a luxury to be enjoyed. While traditional sectors can reinvent the wheel, it defeats the point of implementing a new AI strategy to save time and money if time and money are being wasted in the process. It is much more cost-effective to reach out to others and be willing to interact across verticals and industries in order to get the best solution in place for the problem at hand. Those companies that think outside the box and draw on fresh perspectives, processes and personnel from outside their own sectors are likely to be the ones that will get ahead in the end.