The Second Digitalization: Shedding Light on Your Company's "Dark Data"
Welcome to the first edition of our newsletter dedicated to the second wave of business digitalization. In this issue, the Manufacture Thinking think tank tackles a major but invisible challenge: the "dormant stock" of your information.
Dark matter syndrome
By analogy with the 95% of energy and dark matter that make up the universe, businesses today operate using only 5% of the informational resources at their disposal. While the first wave of digitalization focused on structured data (numbers, names, customer databases), the second is tackling what is known as Dark Data : approximately 95% of the data generated by a company that is never used.
The use of this dormant data by artificial intelligence is precisely what defines the transition to this famous "second digitalization".
What exactly is Dark Data?
What characterizes today's AI is its ability to process unstructured data involving the five senses: sight, sound, smell, touch, and taste.
Dark Data is all the information collected during your daily activities, but which remains stored in invisible corners of your system. A vast field, mostly untouched by interpretation, which includes:
Conversations : internal and external (emails, telephone calls, meetings, instant messaging).
The view : images, photos, videos, older versions of unlabeled documents.
Hearing and smell : like the noise or emissions from your machines.
The IoT : the raw data generated by your various sensors.
How AI turns on the light
AI (and more specifically generative AI and Natural Language Processing) acts as a translation tool, enabling the reading of these formats on a large scale. Here are three direct applications:
Sentiment analysis : AI scans years of emails to identify early signs of customer dissatisfaction or employee resignations.
Predictive Maintenance : By analyzing often ignored machine logs, AI detects micro-variations in temperature or vibration that announce a breakdown months in advance.
Knowledge Extraction (RAG) : AI can read thousands of PDFs of old contracts or technical manuals to instantly answer the complex question of an engineer in the field.
3 "Quick Wins" to move from theory to practice
The central question is: how do we transform this "dead weight" into an engine of growth? The shift from Dark Data to AI is changing the economic model. It's now about creating new sources of profit, not just promises of increased productivity.
Here are three concrete examples of deployment:
The Augmented Customer ( Retail ): A food wholesaler deploys AI agents to check the menus of its restaurant clients online. By cross-referencing this data with purchase history, the company identifies ingredients that the restaurant doesn't buy from them, allowing them to send targeted offers to increase sales.
Smart Metering ( Energy ): Electricity distributors receive a colossal amount of dark data. AI allows them to analyze usage patterns or excesses to identify malfunctions, paving the way for billable on-site investigations. The result? Increased revenue for the distributor and energy savings for the end customer.
Visual Analysis ( Health ): In medicine, patients can take photos of themselves which are instantly analyzed by an AI agent to detect a problem, such as diabetic retinopathy identified via Deep Learning .
Analysis in brief by the think tank Manufacture Thinking
The observation : 95% of the data generated by companies (the "Dark Data") lies dormant in systems without ever being used.
The breakthrough : The second digitalization is defined by the use of AI to process this unstructured data related to the senses (images, sounds, texts, IoT).
The opportunity : Transform this dead weight into a growth driver. The objective is no longer simple productivity, but the creation of new sources of profit through concrete use cases (Quick Wins).
Don't just let the future happen to you. Shape it.
