Automation is the key to unlocking great and lasting advantages in companies in all sectors.
Big data has been nothing short of a strategic automation approach.
On the one hand, we are in an annoying time of information richness, from equipment performance to consumer social media behavior (more than half of all citizens in the world are on social media). But without specific automation (the use of machines and algorithms to manage, process, and analyze available data), your business will lose a great deal of potential potential.
When done well, automation turns big data into “dead” in a living breathing resource that you can use to promote value. So it’s no surprise that many businesses aim to automate anything that can be automated, a Google chief executive recently said.
To help you think about automation in your business context, I present to you three main ways in which this technology-driven activity can help you create value.
It’s the first thing that helps you do automation feature extraction, or extract a critical needle from a large pile of data. Imagine that your organization needs to review patent applications for information about a specific and related technology. You might study thousands or tens of thousands of applications, each with 30 pages or more, for millions and millions of words. But only a small proportion of the relationship between these words and patents matters, for example, dependent on patented technology or the qualifications of inventors and patents of the past.
This task, therefore, like many in the business domain, involves a very low signal-to-noise ratio, and would require thousands of people-hours to complete manually, something that saves too much cost and time. But a machine-based algorithm could be trained to find the key information needed fairly quickly, saving a lot of time and effort. Also, say that in the future you want to search for the same or related set of patents but different information, such as the size of the patent applicant group. You can easily reprogram or reshape the algorithm to take on this task, resulting in economies of scale and higher returns on initial investment.
Second, automation helps data verification and clearing. Data sets often need work. There are errors and missing values, anomalies and sometimes evidence of bias. For example, if an algorithm is trained to detect the characteristics of offenders, but uses only the data of the offenders caught, the algorithm will be sidelined because it lacks data about the offenders who were not caught. to be underestimated. Again, checking and targeting such a large volume of potential problems is too much to take manually. But automation allows for rapid deployment of testing and cleaning tools, while again saving time while creating value.
Third, and this is great, is automation the engine of analytics. Yesterday’s simple regression analysis has become today’s clustered and randomized forest, driven by machine learning, whether to understand product users, optimize inventory for next month’s sales forecast, or predict the impact of a new advertising campaign. Machine-based automation not only repeats standardized low-cost analysis processes on a regular basis, but can also detect nonlinear models that we humans cannot.
For example, my lab analyzed more than 5 million patents using algorithm-driven analysis to see if we could predict the debut of future innovative technologies based on their patent application information. We hypothesized that the machine would identify future success patents from application data if the invention had autonomous and “miracle-like” capabilities or ideas. After all, the algorithm accurately found future successful patents, but not as we humans had imagined. That is, the algorithm did not identify any future success patents based on its autonomous capabilities; on the contrary, they were successful in identifying whether they were part of an a that identified patents cluster among affiliated patents, which can be solved by combining specific problems that an individual patent could not solve on its own.
For example, ultrasound technology has had a profound effect on health for the first time since its first introduction, enabling the treatment of physical conditions such as kidney stones and some cancers and non-invasive imaging. But this advancement would be impossible without a smaller-scale invention beyond basic technology: applicators, static reduction processes, specialized medical cushions and wrenches, developed independently of ultrasonic technology but important for successful application in medicine. Our automated analysis reliably recognized the existence of these related patent sets in more than 5 million patents from healthcare products to the latest golf ball technology, and that these clusters were related to the probability of patents becoming the dominant technology of the future. not previously estimated inference.
My colleague from the Northwest, Andrew Papachristos, used similar analyzes to suggest that Chicago police corruption does not come from a few “bad apples” officers, but from a connected police network that acts in bad faith; his work allows us to detect these issues earlier.
I hope you’ve made it clear that the benefits of automation are mutually reinforcing, and how it can help you turn data into great, sustainable value. In fact, the more data you have, the more automation you need; when you have strong automation capabilities, you can collect and leverage even more data, and the cycle continues.
Conclusion: Automation is an increasingly critical capability, and can be critical to your business’s short- and long-term performance. But it’s important to understand how it promotes value and take steps to alleviate its real disadvantages, for the benefit of your company and the broad community in which you operate.
In the second part of this article, I will discuss the three main disadvantages of automation — scalability, transparency, and cost — and how to address them.