By Redazione LineaEDP15/06/2022Updated:14/06/2022 Facebook Twitter LinkedIn
Fabio Pascali, Regional Director of Cloudera, explains the importance of a data overview for every AI strategy
Fabio Pascali, Cloudera Artificial Intelligence (AI) has already entered our daily lives and is now used by most organizations, including in Italy. Everyone knows driverless cars or voice assistants like Apple's Siri, Microsoft's Cortana or Google's Alexa, but the applications are countless although perhaps less known. There are intelligent algorithms, that is, able to self-learn, which suggest the products to be purchased, movies or music tracks in line with our tastes, know how to answer customer questions via chat, can recognize the face of a person to enable access, sort documents based on content, support doctors in reading x-rays and diagnosis, filter resumes to select the ideal candidate. While there are many applications already known and operational, at least as many are still being tested, which promise to change the way companies operate and individuals manage their lives every day. According to data from the Artificial Intelligence Observatory of the Politecnico di Milano, the Artificial Intelligence market in Italy grew by +27% in 2021, reaching 380 million euros, a value doubled in just two years, for 76% commissioned by Italian companies (290 million euros), for the remaining 24% as exports of projects (90 million euros). The same Observatory has currently identified eight classes of Artificial Intelligence solutions: 1. Autonomous Vehicle: refers to any self-driving vehicle used for transport by road, water or air, such as the self-driving car or the vehicle for home parcel deliveries. 2. Autonomous Robot: robots, more or less anthropomorphic, able to move, manipulate objects and perform actions without human intervention, drawing information from the surrounding environment and adapting to unforeseen or coded events. 3. Intelligent Object: all those objects, from glasses to suitcase, able to perform actions and make decisions without requiring human intervention, interacting with the surrounding environment through sensors (thermometers, video cameras ...) and actuators and learning from the actions of the people who interact with them. 4. Virtual Assistant and Chatbot: the most advanced systems are able to understand the tone and context of the dialogue, store and reuse the information collected and demonstrate resourcefulness during the conversation. These systems are increasingly used as the first level of contact with the customer through the company's Customer Care. 5. Recommendation: these are solutions aimed at addressing preferences, interests, decisions of the user, based on information provided by him, indirectly or directly. Widely used in eCommerce or video and music services (the suggestions of Amazon, Netflix and YouTube are an example), they can be placed at different points in the customer journey or, more generally, in the decision-making process. 6. Image Processing: systems capable of carrying out photo or video analysis for the recognition of people, animals and things present in the image, biometric recognition and, in general, the extraction of information from the image / video. 7. Language Processing: provides language processing skills, for understanding the content, translation, up to the production of texts independently, starting from data or documents provided as input. 8. Intelligent Data Processing: all solutions that use artificial intelligence algorithms on structured and non-structured data to extract information: for example, systems for detecting financial fraud, pattern search, monitoring and control systems, predictive analysis. It is easy to understand how in each of these categories the advantages compared to any manual process are potentially infinite, in terms of processing capacity, speed and accuracy of the processes. Only Artificial Intelligence can in fact collect, catalog and analyze the large amounts of data that come from the systems and applications in use, and return in a short time results that can be used for business purposes. In the specific business environment, the use of AI solutions can be traced back to three fundamental macro areas: the improvement of decision-making processes, the innovation of products and services, the reduction of costs. The Pwc 2021 AI Predictions research has clearly shown how companies that have started to make more advanced use of Artificial Intelligence find obvious and measurable advantages in these areas.
With the right data and models, AI enables better products, greater productivity, and a more rewarding customer experience, which can increase the number of projects and customers, resulting in more data available, in a virtuous circle that results in even better products and experiences. Among the most popular use cases are those related to reducing costs through automation and improving low-value, high-cost processes. This is the case, for example, of customer relations solutions or those aimed at increasing productivity in manufacturing companies. The other area of more concrete current use concerns the application of AI-based tools to improve customer understanding, reduce churn in service companies, or understand how to quickly increase sales in a particular industry. However, it will be decision-making processes that will benefit from AI even more importantly: organizations struggle to make good decisions, especially in contexts where the specific process draws advantage from almost immediate response times. Thanks to intelligent decision making applications, managers and employees will be able to make better decisions, in a shorter time. Artificial Intelligence, however, is not an element that lives for itself, but is part of a broader data strategy, which can bring its benefits the more it includes all the business applications called to manage the various aspects of the data cycle itself. And it is precisely this cycle that must be considered as a whole, also from the point of view of AI. If you focus on the realization of a single project you will certainly solve the business problem associated with that specific application, but you will lose the opportunity to use the data processed by that AI application for other purposes as well. And this is the most delicate step: moving from a realization of vertical AI projects to an industrialization of AI that is no longer based on the single silo of information, but that collects it from a corporate Data Platform. In this way, the data used in real time by an AI application , appropriately processed by Data Engineering tools and effectively moved to advanced Data Warehouses, Data Lakes or Operational Data Bases, can be made available to other teams of Data Scientists in order to develop new ML and AI algorithms that can generate valuable insights for business success.Curated by Fabio Pascali, Regional Director of Cloudera