A five-phase method that treats artificial intelligence the way a board treats any investment: with a defined perimeter, an expected return, a working pilot, verified results — and permanent human-in-the-loop oversight.
Most AI programmes start with tools and end with slideware. PROVA starts with the business case and ends with measured, governed adoption — in that order, every time.
Assessment of where AI can operate safely in your business: data, processes, people — and the governance boundaries set by the EU AI Act and GDPR.
A business case before any build: expected hours saved, error rates reduced, time-to-market gained. If the return isn't there, we stop here — cheaply.
One prioritised use case, implemented as a working pilot in your real workflow. Not a demo — a tool your team actually uses, with human oversight built in.
Systematic evaluation against the baseline: before/after metrics, audit trail, documented human-in-the-loop. Results a CFO — or a regulator — can trust.
Scale what is proven: training for your team, an internal governance framework, and a roadmap for the next use cases — owned by you, not by a vendor.
Thirty years of corporate finance — IPOs, M&A, restructurings — mean every recommendation is framed the way owners and boards decide: outcome, risk, return on capital.
AI Act, GDPR, DORA: compliance is designed into the perimeter from day one, not retrofitted after the first incident. Built for businesses that answer to regulators, clients and auditors.
No transformation theatre. Every engagement produces measured results on your own operations — and you keep the assets: playbooks, tools, trained people.
A senior corporate finance professional who works with AI daily — and teaches executives and their teams to do the same, safely.
Francesco has spent three decades advising Italian companies on the most consequential decisions of their life: going public, buying, selling, restructuring. Today he combines that judgement with hands-on fluency in applied AI — building agents, knowledge systems and evaluation frameworks, not just talking about them — to help European SMEs, professional firms and family offices adopt AI with the discipline of an investment committee.
"AI adoption is not an experiment to delegate: it is an investment to govern with numbers and defend with proof."
Milan, Italy · European Reach
From an initial executive diagnostic to full operational implementation and continuous board-level oversight.
A focused session with the owner or board: where AI moves margin and risk in your business — mapped in 90 minutes, in your language, on your numbers.
Request Briefing →The full P and R phases: use-case discovery, prioritisation by impact and feasibility, business case with expected ROI, and a governance perimeter ready for the AI Act.
Request Assessment →Ongoing oversight as your AI footprint grows: quarterly steering, compliance watch (AI Act, GDPR, DORA), and independent verification of what your vendors promise.
Retain Advisor →Empowering professional firms and corporate teams to build, govern, and deploy bespoke AI tools — without writing code manually.
Hands-on programs designed for commercialisti, lawyers, CFOs, and executives to master Claude Code and build internal software tools.
Complete operational enablement for professional firms and finance teams: secure local terminal setup, balance sheet reclassification, and contract due diligence.
Workshop Details & Syllabus →Build bespoke web apps, financial runway dashboards, and automated document extractors without writing code manually. Driven by Claude Code & Antigravity under Specification-Driven Development.
Apply for Next Cohort →Private bespoke mentoring for Managing Partners, CFOs, and Board Directors: installation on your personal workstation, custom decision-support instructions, NDA protected.
Inquire for Private Coaching →Four recent engagements, anonymised for confidentiality. Every figure below is either measured on the client's own operations — from the actual work logs — or declared as a baseline in the approved programme.
The Context. Due diligence on over 1,000 potential partners and verifying compliance documentation across 15 European supplier portals previously required weeks of manual desk-work.
PROVA Engagement. 20-hour executive programme for the two owners, deploying parallel AI agents with explicit exclusion criteria, human oversight gates, and customs intelligence for a board-ready report.
The Context. The group wanted to test AI beyond vendor slideware, aligning strategy, IT, after-sales service, and marketing without dispersals or vendor lock-in.
PROVA Engagement. Grown from a single pilot module into a 4-function corporate programme. Handed over 9 reusable custom AI assets (skills and prompt libraries) owned entirely by the client.
The Context. Increasing document complexity required streamlining legal research, memos, and due diligence without compromising professional secrecy, GDPR, or deontological verification.
PROVA Engagement. A first training phase for the firm's lawyers, in groups, working only on public or anonymised data, with 15 reusable skills codified on the firm's house style.
The Context. Managing family assets, price-sensitive listed-company information, and payment fraud risks, with the need to satisfy EU AI Act Article 4 literacy obligations.
PROVA Engagement. 12-month, 135-hour organizational programme across 6 functional areas. Delivered AI Usage Policy and Integrated IT+AI Policy before day one, cutting fund due-diligence cycle in half.
Methodology: effort figures are measured from the file and session logs of the actual engagements; manual baselines are declared estimates based on market practice. For programmes currently in delivery, scope figures and KPI baselines reflect the approved programme documents. Client names are withheld for confidentiality.
People are already pasting company data into public chatbots — no perimeter, no log, no accountability. Prohibition doesn't work: the answer is a governed alternative that beats the workaround. This is the regulatory backbone built into every PROVA engagement.
AI literacy (art. 4) is already a legal obligation, and the transparency duties of art. 50 apply from 2 August 2026. A structured training programme is not just capability building — it is the evidence a regulator asks for.
A clear data perimeter before the first prompt: what may enter which tool, in which environment. Public and anonymised data first; enterprise environments with data-processing agreements before any confidential file. Client files and family data treated as a security matter, not a formality.
Finance and listed-company ecosystems answer to rules of their own: price-sensitive information and market abuse, DORA for financial entities, Italy's AI law 132/2025. If your business answers to a regulator, your AI adoption must answer to it too.
A register of authorised tools that says yes fast — slow approval is what creates shadow AI. Human-in-the-loop checkpoints where errors are expensive, an audit trail where decisions matter, and a policy your people can actually follow.
Take our 16-point interactive self-assessment on EU AI Act Article 50 and Italian Law 132/2025. Instant score, no email required.
The first conversation is a working session, not a sales call: bring one process that eats your team's hours, and leave with a view on whether AI can pay for itself there.
Reserve a 30-minute private consultation with Francesco:
Book 30-min Executive Briefing →