The first digitization: necessary, but incomplete.
Most SME leaders give the same response when asked about digital transformation: “We’ve already done that.” Modernized ERP systems, digitized processes, CRM in place. The box is checked. It is precisely this certainty that deserves to be questioned today.
Since the 1960s, digitization has focused on using structured data to support day-to-day operations: customer names, product codes, and financial figures. Data that fits neatly into a spreadsheet and integrates into an ERP system. This model has delivered real productivity gains, but it has imposed a very narrow definition of what constitutes “useful data.” And this definition has left out the essentials.
Insight: Written by Delphine Seitiee, Secretary General of Alp ICT; Xavier Comtesse, Co-founder of ManufactureThinking; and Cédric Fischer, Partner and expert in communications and public relations at Index
95% of your data goes unused
In a company, approximately 95% of the data generated each day falls outside the scope of traditional systems. This is unstructured data: email exchanges with your customers, meeting minutes, call recordings, photos of technical service calls, equipment alerts, and contract archives. All of this exists within your organization. All of this is generated, stored, and largely overlooked.
We call this “dark data”: a dormant pool of information that your company generates but doesn’t use. The comparison with the universe is telling: visible matter accounts for only about 5% of what exists; the rest eludes our instruments. The same is true within your organizations. You’re running your business based on just 5% of what it knows.
What AI Changes: Reading What Was Once Unreadable
What makes this moment strategically unique is that AI can now “read” this unstructured data, interpret it, and extract meaning from it—quickly, at scale, and at a cost that’s affordable for small and medium-sized businesses. This is no longer just research or prototyping. It’s available now to organizations of any size.
A service provider can automatically analyze years of customer interactions to detect subtle signs of dissatisfaction—well before a contract is terminated. A manufacturer can transform its equipment logs, which were previously archived solely to comply with regulations, into a system for issuing preventive alerts about potential breakdowns. An engineering firm can convert its field documentation archives into a tool for remote condition analysis and maintenance recommendations.
In each of these cases, the asset already existed. What was missing was the ability to read it.
From Cost to Profit: A Shift in Perspective
That is where the real paradigm shift lies…
To read the rest of this article and discover many other articles on the topic, simply sign up for the Observatoire du Numérique, a platform for continuous monitoring of information and digital technologies developed by Alp ICT. It’s free!
