By Bruno Maia *
Digital Transformation, Internet of Things, Artificial Intelligence, Big Data. In recent years, these terms have been highlighted and discussed in profusion by professionals from different industries. Companies are increasingly aware of the importance of innovating in order to remain competitive and relevant, especially in an extremely competitive scenario and susceptible to the rapid changes that the Digital Age imposes.
Thus, I would like to address in this article a technology that has the potential to profoundly impact business in the very near future: Computer Vision. It is a branch of computer science that uses Artificial Intelligence to identify and process images in order to generate data that can be analyzed and used to assist in making important decisions. This niche in the Advanced Analytics market has been gaining relevance over the years. According to Market Research Future, Computer Vision had a value of just over US$ 9 billion in 2017 - the forecast is that the value will increase to US$ 48 billion by 2023 (quite a growth)!
This technology takes Analytics as we know it, by which historical data is used, to make predictions and prescriptions, to another level. Computer Vision gains special relevance when used as a device to produce context in real time to obtain analytical intelligence and achieve much higher commercial optimization rates. Result? It is becoming increasingly relevant and, very soon, indispensable for any company.
To illustrate the potential of applying this technology, combined with Customer Intelligence, I like to use the Retail segment as an example. Typically, websites use benchmarks for converting shopping carts that are based on data from connected customers, such as purchase history. Computer Vision allows the context to be used to predict, prescribe and adjust which initiatives are necessary to be adopted to dramatically increase the chances of conversion of that client. A beauty products company can launch an application that captures the customer's face through the cell phone camera to estimate their age, ethnicity, skin type, among other factors. In this way, they can suggest a set of makeup or other products that have more adherence to the personal characteristics of the consumer.
The technology can also be used to measure, by means of cameras, the number of customers of a certain store, define the heat map, the organization of the shelves, among other applications. What's new here is the fact that the cost of this equipment has fallen a lot, its precision and resilience increased and also the possibility of applying sophisticated algorithms in real time with the aid of on-edge computing, allowing it to be necessary to be connected in the cloud to apply the concept. By combining all customer insights it is possible to achieve a conversion rate of 17% - the usual revolves around 5%. This is an exponential growth, which can take business to another level!
The evolution of Visual Computational - whose roots go back to the concept of Optical Character Recognition (OCR) - has also evolved significantly. Currently, it is already possible to mix contexts in real time and that can come from different sources, such as cell phones, Wi-Fi, camera images, in places such as shopping malls, and even fiber optic networks. It is possible to use these sources and a wide range of technologies to capture aspects of the environment, classify data and, thus, create and train increasingly accurate models. Governments have even used technology to increase road safety, for example, by monitoring the safety of access roads and even assessing bridge wear in real time.
With the continuous evolution and improvement of new technologies, industries have gained powerful tools to predict the future, in order to offer increasingly personalized and assertive experiences to their customers. Whoever manages to innovate more quickly will certainly take the lead in the race - increasingly fierce - for consumer preference.
* Bruno Maia is Head of Innovation, SAS Latin America.
Warning: The opinion presented in this article is the responsibility of its author and not of ABES - Brazilian Association of Software Companies
EN 
PT 
