Insights Abacus

AI in ICT: bringing it into production, not just experimenting with it

Home / Insights / AI in ICT: bringing it into production, not just experimenting with it
On this page

Introduction

Artificial intelligence is no longer an experiment.

In Italy, it reached €1.2 billion in 2024, growing 58 percent year on year.

But the point is no longer to adopt it.
The point is to actually make it work in business systems.

Many initiatives stop after the go-live.
Few get to scale, integration and business continuity.

The context: what’s going on

AI adoption in the Italian market is accelerating rapidly:

  • 59% of large companies have at least one active project
  • Adoption is driven by sectors such as finance, energy, and utilities
  • SMEs are increasing investment, but at a more gradual pace

The result is a two-speed system, where the ability to execute makes all the difference.

Where problems arise

The main critical issues are not technological, but operational:

  • Scalability – Taking AI from project to system
  • ROI – demonstrate concrete impacts over time
  • Compliance – managing AI Act, GDPR and governance
  • Skills – Spreading real use of AI in organizations

The risk is clear: AI adopted but not integrated.

AI nell ITC

Where you are investing

The highest value areas are:

  • Process automation and orchestration
  • Predictive analytics and decision support
  • Generative and conversational AI
  • Vertical applications by industry

New focuses are also emerging:

  • Agent-based AI to orchestrate complex systems
  • AI reliable and compliant
  • AI for sustainability and ESG
AI nell ITC

What really works

Companies that succeed in scaling AI take a different approach:

  • They integrate AI into existing systems
  • They build robust data architectures
  • They manage patterns over time (MLOps / AIOps)
  • They address governance and compliance from the very beginning

In short, they treat AI as part of the infrastructure.

AI nell ITC

Abacus approach

Abacus works on AI as a component of the IT backbone.

Key areas:

  • Generative & Predictive AI
    for decision-making scenarios and simulations
  • Conversational & Agentic AI
    To manage complex interactions and processes
  • Process orchestration
    to govern end-to-end workflows
  • Responsible AI
    to ensure compliance, transparency and trust

Expected Results.

A structured approach to AI enables:

  • Increased scalability and operational resilience
  • Faster and more informed decisions
  • Regulatory alignment and effective governance
  • Increased confidence in the use of AI
AI nell ITC

Case studies

AI nell ITC

Background:
Global automotive company with multiple manufacturing plants

Challenge:
Implement a scalable and integrated predictive maintenance system

Solution:

  • Integrated data platform and AI
  • MLOps / AIOps Architecture
  • Agent AI modules for process orchestration

Results:

  • 30% reduction in false alarms
  • Failure advance over 7 days
  • 40% reduction in unplanned downtime

Download

Download the full version of the insight.

AI in ICT: bringing it into production, not just experimenting with it

Full Insight

Download PDF
RELATED

Related Insight

Learn about other Abacus insights related to the topics of AI, ICT and business technology evolution.

Cybersecurity
in the age of IoT

From exponential device growth to new attack surfaces: how to build security, governance, and operational resilience.

Cloud
and Edge

Scalability, real-time and control: because cloud and edge are not alternatives, but components of a single operational architecture.

Customer Experience
and AI

How AI is transforming customer management and what it really takes to scale personalized, reliable and sustainable experiences.

Want to understand how to bring AI into production in your context?

Let’s talk about your needs and build a concrete approach.

Contact us