Business in the data world: the oracular economy

Interview with Cosimo Accoto, research affiliates @MIT


Among lines of data and their infinite recombination, a new digital world is taking shape. What technological impacts and strategic repercussions is it going to bring for the development of innovative businesses? Let’s discuss it with an expert of the most advanced frontiers of digital thought and philosophy, Dr. Cosimo Accoto: a philosopher of training and research affiliated with MIT, author of the essays "Il mondo ex machina" (Egea 2019) and "Il mondo dato" (Egea 2017).


A: What do you mean with extension of the data domain? And how could data collection and analysis technologies become "oracles" for business?

Q: I mean the growing importance of data in business, on many dimensions and perspectives. We can mention the progressive process of "datafication" (data reduction) that markets and companies are experiencing. The forms of this quantification are multiplying to the point that today we can say that every business is, actually, a "data business" whatever the product, service, experience that companies proposes to the market.

This extension of the data domain concerns different strategies:

  • the ability to monitor and quantify both natively digital platforms and sensorized physical environments (with the so-called digital twins, for example, and not only for digital businesses),
  • the ability to support business intelligence in a broader sense (from data-driven innovation to the analysis of the relationship with consumers),
  • the ability to process the data in a multidimensional way (not only for descriptive and diagnostic purposes, but also predictive and anticipatory, prescriptive and decisional purposes).

Therefore, data has grown to a dominance that cannot be canceled. Moreover, the power of data impacts on three dimensions: operational (it is used to optimize operational processes), transformational (it is central to innovate business models), institutional (in perspective, it unhinges and redesigns eco-systemic relationships, supply chains, product sectors and markets). As I wrote in my essay "The data world" (Egea), with data the most important transformation is the passage from "archival" to "oracular" economies.

As a matter of fact, the new AI technologies are building a technological-informatics architecture (as well as an economic and a business one) in which information begins to flow, systematically, from the future to the present, and no longer from the past to the present as it has been until now. By using sensors, data and algorithms of artificial intelligence, machines are able to intercept information about what is about to happen, and use this probabilistic information to design services and products in advance (and not just in a postponed and responsive way).

For example, we are accustomed to the shopping-then-shipping model (buy first, then I'll send it to you) - write the authors of the recent essay "Prediction Machines". But we could soon redesign and reverse the experience: shipping-then-shopping (I'll send it to you, as I know you’ll buy it) because predictive analysis can accurately anticipate what consumers are going to buy in the future. This will apply to many sectors, from preventive medicine to predictive maintenance, and so on.


A: Could applying Artificial Intelligence to data help make Marketing Automation more effective? How can we customize more with less human effort?

Q: Data was already relevant, but with this new spring of artificial intelligence and the success of machine and deep learning it becomes strategic. That's why, as Andrew Ng recently said, every artificial intelligence business strategy is primarily a "data strategy".

Data today is derived from the sensors of the Internet of things (for example, for maintenance or connection between machines), from human behavior on social media (especially photos, videos, texts, clicks), from logistical movements and mobile devices (calls, urban routes, tracking of health or sports performance with wearable accessories) and so on.

After a first phase of this data revolution, we’ve entered a second phase where data fuels artificial intelligence. In fact, what we are experiencing today is strongly driven by the ability of machines to learn from experience and data. The successes of self-driven vehicles, for example, are the result of machine learning and deep learning techniques that feed on data.


Through the training of artificial neural networks (convolutive and recursive), the "intelligent" cars experience and perceive the world, extrapolate models and knowledge, interacting with the environment in an increasingly sophisticated and autonomous way. The very same logic and techniques are increasingly applied to marketing automation, by exploiting personal data, physiological sensors, customization algorithms and social platforms. In this way, custom content and experiences can be created in an automated way, in near-time (not only in real-time), and in a pervasive and granular way.

The process is well explained in a recently published book "The Invisible Brand. Marketing in the Age of Automation, Big Data & Machine Learning". There has been, undoubtedly, an optimization of the human intervention (many of these operations are machine-based), but what I think it’s more interesting are the chances for innovation related to this new intelligent automation. We will have to face, of course, critical issues concerning privacy (past/present data) and destiny (present/future data), but the ability of machines and brands to know consumers will allow us to ultra-customize experiences, content, services.

In this sense, in my latest essay "The ex machina world" (Egea) I suggested moving from the concept of touch point to that of data switch: every occasion of interaction between brands and consumers is not a passive point of contact (as we call it today), but a moment of data exchange that allows a dynamic and anticipated customization.

A: In your opinion, what are the main steps to be taken in order to become a data-driven enterprise?

Q: Digital companies are facilitated in this. They have a native data culture because it is directly rooted into their business. Often, in digital companies, data camps are held for all employees (not just for analysis departments) to support and make palpable the importance of a data-centric approach, or one that is strongly connected to data. 

This is more difficult, however, for traditional companies, which do not consider data as a key asset, but as separate from operations and generally relegated to the department of analytics and business intelligence. Here, data is used to periodically monitor trends and for market research, and occasionally for marketing, but it is not yet considered the lifeblood of the entire organization and production. I believe that there’s primarily a problem of a lack of "data culture" and of its relevance. Much is decided on the basis of feeling or past experience alone, and data is often not updated or not shared across the departments. 

As I anticipated, there is a need to design and implement a data strategy with an agile and adaptive approach, favoring the ability to learn more than sticking to plans or performance. In addition, it’s important to define the key needs and business objectives beforehand, and then select platforms, processes, resources and skills accordingly. It can mean, perhaps (if you own and sell physical products or physical infrastructure) to imagine a digital twin or sensorized digital twin of these that is able to collect information on product use or performance of the structure. 

Let’s make a point: it is not simply a matter of having a data platform within the company (even if you buy adequate hardware and software). It is about being a data platform, of companies becoming organizations that can feel their markets and consumers, collect and process information, scale their ability to learn and consequently build experiences, applications, services and products. 

All of this with respect, of course, for the consumer’s privacy and ethical principles. And in the future, we will probably also have data management forms that will make consumers and users protagonists and no longer, as happens today, just passive or unconscious (and, sometimes, abused) objects of measurement.