AI training for the 16 managers of Farvima MedicinaliThe pharma distributor with over a billion in revenue and 800 employees
01 · In short
From problem to result
The problem
They were not starting from zero: everyone already used an assistant. Nothing stuck, though. The same data retyped by hand into the same format, every week.
The solution
An assessment filled in by the whole group before we walked in, then 2 days on site with Claude Code. Each manager started from a template customised on those answers.
The result
16 managers, at least three skills each. They automated part of their management control, financial statement analysis and debt collection, plus dashboards on sales and competitors. In 2 days.
02 · Who Farvima Medicinali is
Farvima Medicinali is one of the largest pharmaceutical wholesalers in Italy, based in Casandrino, near Naples.
The company in six numbers
12,000 pharmacies served
16 warehouses
95,000 product lines in the catalogue
7.3 million deliveries a year
800 employees
Over €1 billion in revenue
The business is demand forecasting, stock and deliveries to pharmacies. A check done by hand multiplies across tens of thousands of product lines and across every day of the year.
This is not a company new to automation. The physical warehouses have been automated for years. That is precisely why the bottleneck had moved somewhere else: onto whatever still sits in people's hands.
03 · Why we started with the 16 managers
In the room were the 16 managers who run the company's functions, from administration and sales through to IT. Their so-called "AI Champions".
In a group serving 12,000 pharmacies, a check done by hand by one manager is not that manager's private problem. It is a procedure repeated across their whole function, every week, for years.
That is why we started with them. They are the people who design how everyone else works.
04 · The assessment, before we walked in
As in every training programme we run, before we walked in we had the whole group fill in an assessment. We use it to find out where each person's time really goes: task by task, with which tools and how often.
« Without that measurement you teach AI in general. And AI in general automates nothing. With it, every person already knows walking into the room which of their own tasks are about to disappear »
What came out
Nobody was sceptical
Everyone already used an assistant. Asked about their main blocker, the largest group said they did not have one: they were already using it.
The real blocker was elsewhere
The same data from different sources, retyped by hand into the same format: control reports, reconciliations, price list comparisons, pulling figures out of shipping documents. More than half the group reported at least one task taking over five hours a week.
05 · How we built the days
The tool is Claude Code. Nobody started from a generic example. From that template onwards, what each manager builds is theirs and stays theirs.
The four steps
The assessment
Before the room. Where each person's time goes, task by task.
One template each
We built it on that manager's role, customised on their constraints and the documents they touch.
An hour and a half of theory
The minimum needed to grasp what a skill is and why a model needs context.
All the rest hands-on
On each person's own computer and their own real processes, not on a sample.
06 · What they built
With their own hands, on their own processes
A competitor analysis dashboard
Reporting analysis
Supplier deadline management
Financial statement analysis
A sales control dashboard
Management control
Debt collection
The scale
16 managers trained in two days
At least three skills each, built with their own hands
We did not choose any of these tasks. They were all on the list that came out of the assessment, written by the people who do the work.
« I was extremely satisfied. We got more done today than in many months »
Mirko Di Falco, CEO of Farvima Medicinali
07 · How we carry on
The days do not close the work with Farvima: they open it. We carry on across the rest of the management team, function by function, until all their key users are covered.
"Ha smussato quello che per molti imprenditori, e lo è sicuramente per me, è la montagna dell'AI: me l'ha resa una collinetta da poter inserire in azienda. Perché ovviamente tutti noi usiamo l'AI e varie piattaforme, ma l'utilizzo aziendale che porta valore è tutta un'altra storia. Sono stato estremamente soddisfatto. Abbiamo fatto più oggi che in tanti mesi."
Want similar results for your company?
Tell us about your challenge. We'll find the right AI solution for your business together.
Let's Talk