Applied artificial intelligence
AI that works inside your processes. Not inside a slide deck.
76% of Italian SMEs have neither invested in AI nor plan to. Not out of laziness: because nobody has yet explained what actually changes on Monday morning. That is where I start.
Applied business AI is the use of language models and automation to handle repetitive tasks that currently occupy people: reading documents, extracting data, drafting standard replies, classifying requests, preparing quotes. It does not replace judgement — it removes the mechanical work around it.
Where you actually start
You start by counting hours. Before proposing any technology, I measure how much time your team spends each week on repetitive work. If there are not enough hours to recover to justify the investment, I say so and we stop there.
The candidates are almost always the same: sorting incoming email, retyping data from a PDF into a system, preparing recurring quotes, answering questions already answered a hundred times, checking that a document is complete.
Nobody takes pride in this work. Removing it does not make anyone redundant: it gives people back the hours for the part of the job that actually needs a brain.
The question of control
41% of executives say they fear a lack of control over AI decisions. That is a legitimate concern and it is solved in design: every automation I build has a point where a person can see what happened, and a point where they can stop it.
A rule I do not negotiate: AI does not make decisions with direct legal or financial consequences without a person confirming them. It prepares, proposes, sorts, flags. The signature stays human.
Every automation keeps a record of what it did and why. If something goes wrong one day, it can be reconstructed. A system you cannot inspect is not a system — it is a liability.
And your data?
Your data stays yours. I use configurations where company content is not used to train third-party models, and when data is sensitive I evaluate solutions running on your own infrastructure.
67% of executives believe their company has already suffered a data leak through unapproved AI tools used by employees. The problem is not AI: it is AI used in secret because the company never offered an official alternative. Providing a safe tool is also the most effective security measure available.
What it includes
- Start from repetitive work, not from the technology
- Automating documents, email, quotes, classification
- Internal assistants trained on your own company documents
- Your data stays yours: never used to train third-party models
The other routes
01
Custom management software
A system built around the way you actually work, not around the way an off-the-shelf product insists you should.
Read more03
Websites and e-commerce
A site that brings in enquiries while you do something else, and that gets found both by Google and by the AI assistants people now ask for advice.
Read moreThe first step
We start with a conversation, not with a quote.
Forty-five minutes in which you tell me where your day gets stuck. At the end you will have at least one useful thing: my honest opinion on what to tackle first — even if you then decide to do it with someone else, or not at all.
No commitment, no email sequence afterwards. If you are not the right client, I will be the one to say so.