Pietro Ciattaglia
CEO & Co-Founder
Profile
- Role
- CEO & Co-Founder
- pietro@zendata.it
- About
- Takes AI agents to production on real business processes, from first use case to go-live.
Personal LinkedIn profile: to be added once confirmed (TODO: first-party evidence).
Published articles
Italian AI Law 132/2025 article by article: commentary for companies (2026)
All 28 articles of Italy’s AI law explained one by one: what they say, who they affect, what changes for a company. Including the 2026 implementing decrees.
AI agent interoperability: MCP and A2A explained for companies (2026)
MCP connects an agent to tools and data, A2A connects agents to each other. Differences, what changed in 2026, when an SME needs them, and the security risks to manage.
UNI/PdR 125:2022 certification: a guide for companies in Italy (2026)
What Italy’s UNI/PdR 125 gender equality certification is, who can get it, the 6 areas and 33 KPIs, anonymous reporting, costs, timing and benefits. From a certified company.
AI course creator tools 2026: create online courses from documents
Coursebox, Mindsmith, Easygenerator, Synthesia, Sana Learn, LearnWorlds and EduAI compared by the job they do after a document becomes a course.
Transition 5.0 in 2026: does company AI qualify? (hyper-depreciation)
August 2026: Transition 5.0 tax credit closed; hyper-depreciation remains. ChatGPT and an admin agent usually do not qualify. Check before you buy.
How to connect an AI agent to an Italian SME accounting system (2026)
For owners who run TeamSystem, Zucchetti, SAP, or a CRM and hear “we’ll hook up the AI”. Three honest ways, what the agent may do alone, and why the quote jumps.
Microsoft Copilot for companies: when it is enough and when you need a custom agent (2026)
Copilot Chat, Microsoft 365 Copilot, and Copilot Studio for SMEs. Italy list prices as of August 2026, vs ChatGPT, vs a custom ERP/process agent.
Document dematerialization in Italian companies: compliant archive, DMS, and AI
Digitizing files is not enough for AI. Dematerialization, legally compliant preservation, DMS, and a queryable knowledge base are three different layers — with AgID, CAD, and Istat sources.
ChatGPT Business: vs Plus vs Enterprise for an SME (2026)
ChatGPT Business is not Plus for the team. Price, training, SSO; when the plan is enough and when you need an agent on processes. Check OpenAI checkout.
AI agents for companies: what they are, examples and cost (2026)
An AI agent pursues a goal and can act on your systems. Concrete examples, cost (€15–35k in production), agent vs assistant, chatbot and RPA, authorization.
AI usage in Italy (2026): what real usage data shows (Anthropic Economic Index)
AI usage in Italy tracked from real Claude.ai conversations: 42.9% is work, above Germany, France and the UK. Charts, tasks and limits — not a survey.
How much a custom AI agent costs for an Italian SME in 2026
Realistic 2026 bands: PoC, production assistant, agent with integrations. What drives price, run costs, and questions to ask before you sign.
AI Act Article 50: chatbot and AI-content transparency duties from 2 August 2026
What SMEs must do on chatbots, deepfakes and public texts. Internal ChatGPT for email usually not; a site widget yes. Provider vs deployer.
AI Act: are you a deployer or a provider? SME checklist
Copilot at work vs a branded chatbot: the role changes the duties. Eight points to classify systems, transparency, and when an SME becomes a provider.
Mandatory AI Act training (Art. 4): how to do it in the company from 2025
AI literacy has applied since 2 February 2025 for professional use. Not a master’s: attendance record, a few hours on real work, data rules. How to evidence it.
Law 132/2025 does not replace the AI Act — what companies should do
Italian Law 132/2025 in force from 10 October 2025. It is not a second AI Act: worker notice, AgID and ACN. What changes for an SME.
AI Act: what companies must do from 2 August 2026
The AI Act is the EU AI regulation. From 2 August 2026: inventory, chatbot, literacy. You are not a provider. High-risk duties move to 2027.
AI training in the company without a huge budget: practical literacy for SMEs
Skip generic webinars. Short sessions on real work, data rules and an internal champion. How to get tools actually used.
Company knowledge for AI: make it usable before you pick a model
The strategic question is not which LLM to buy. It is which knowledge to make readable, current and citable. An SME checklist.
Build, buy or partner for AI in an SME: how to choose without wasting budget
Build vs buy vs partner is not a fashion: it depends on process, data and ownership. A framework before you sign a contract.
Where to start with AI in a company: the first real project for an SME
Do not start from the tech: start from the problem. How to pick a measurable use case, with an owner and data, without a 12-month programme.
Is your data ready for AI? A checklist for companies
Many AI projects stall because data and knowledge are not ready. How to judge quality, access, ownership and what to fix first.
AI and work in the company: how tasks, roles and skills change
AI does not automatically remove people: it changes tasks and accountability. How to communicate the shift and train daily users.
Who is accountable when AI is wrong at work? Ownership in the company
AI has no legal personhood: the organisation does. How to name owners, escalation and a paper trail before the incident.
Shadow AI at work: why banning ChatGPT is not enough (and what to do)
Blocking tools does not remove the need: it pushes personal accounts and data out of sight. How to build a usable company path with logging and data rules.
AI agents vs chatbots at work: difference, when they pay off, risks
An agent is not a smarter chatbot: it pursues a goal and uses tools. When it makes sense for an SME, what must exist first, and why so many projects stay in pilot.
Why AI gets it wrong at work: context and documents (not just the model)
Plausible but wrong answers often come from old versions or incomplete sources. How to cut errors by organising knowledge, not only prompts.
How to measure AI ROI in a company: metrics beyond time saved
Recovered time is not enough to know if an AI project works. A pragmatic framework: baseline, quality, adoption and value for SMEs.
Document management and AI: why messy files make the model fail
Company AI runs on documents, not only on the model. How to start from one process, versions and access — before buying another tool.
From AI pilot to production: why projects stall (and how to get out)
Most pilots never reach production. Typical causes (data, ownership, measurement) and how to design the handoff in an SME.