What is digital transformation?

Digital transformation is the change in how a company creates value, competitively or organizationally, made possible by digital technologies: not buying software, but reworking processes and decisions around what that software makes possible. It is distinct from two terms it gets confused with. Digitization is converting information from analog to digital form: a paper receipt becoming a database record. Digitalization is using that digital data to automate an activity that used to require manual steps: an order that used to travel by email and now runs through a system. Neither changes how the company decides or competes: that is the level digital transformation operates on, and it is also why a company can have digitized everything and still have transformed nothing, still be measuring its own progress by the wrong yardstick, deployment instead of decisions genuinely changed.

See also: Change management (technical) · The method for getting a new system or automation actually adopted, not just deployed.

Where Italian SMEs stand

Italy's national statistics office: in 2025 almost 80% of Italian companies with at least 10 employees reach at least a basic level of digitalization (at least 4 of 12 digital activities in the DII index), but only 38.1% reach a high or very high level. On AI the gap is widening: 15.7% of small companies use it against 53.1% of large ones, and the distance between SMEs and large companies, around 20 points in 2023, widened to 37 points in 2025 (ISTAT, Imprese e ICT 2025). The European Commission's 2025 Digital Decade report on Italy measures the same distance at EU scale: 70.2% of Italian SMEs have at least a basic digital intensity, against an EU target of 90% by 2030. What matters most is the gap between intermediate or advanced cloud, adopted by 68.1% of companies, and AI: that gap, more than the average, says how much work remains.

Why projects fail: starting from the tool

The most common failure is the sequence reversed. A project starts by picking a CRM, an ERP or a generative AI platform, and only afterward does anyone clarify which decision, made today on incomplete or late data, that software is supposed to improve. The typical result is a system used alongside Excel: adoption fails because nobody redesigned the process that software was supposed to replace. It is the same pattern described under change management: shipping a system and getting it genuinely adopted are two different jobs, and the second one is usually skipped.

A concrete example

A 40-person distributor of industrial components decides to digitize its warehouse: barcodes, a system replacing Excel sheets, real-time stock levels. Six months later inventory is accurate, but order fulfillment times have not moved, because nobody touched the downstream process: a clerk still prints the order and calls the supplier by hand, exactly as before. The digitized part worked; the transformation, which would have meant rethinking who decides what across the order-to-supplier chain, never started.

Why it matters for decision makers

A pragmatic sequence for an SME starts with a diagnosis, not a quote: where today a business decision rests on incomplete data, late data, or knowledge held in one person's head. From there, picking a process, not a department, measurable in months rather than years, is what the first digital intervention is built around. The right starting point is almost always a minimal data strategy, because without reliable data every automation built on top inherits the same error it had before, only faster; and an honest maturity assessment before any project with AI in it, because the gap ISTAT measures between cloud and AI adoption is not an accident: AI presupposes data already in order, and skipping that step is the most common way to accumulate technical debt instead of reducing it.