From conversation to requirement
What we agree on becomes a user story, a page in a shared knowledge base — not a note nobody finds again.
We help companies turn slow processes, repetitive work and scattered knowledge into measurable value — then help their teams keep building on it. Discovery, delivery and governance, on tools we built ourselves.
Creatyve Studio is where companies come to decide what to do about AI — and to build it. Luca Cagnassi leads content, brand and communication; Sergio Sentinelli leads process, architecture and implementation. We didn't add AI to a list of services — we rebuilt how we work around it, and turned that into the tools we use with our clients every day.
When a company has to decide something about AI — whether it's worth it, where to start, how to keep it under control, how to build on it — we're the ones to call.
Not because we have opinions about AI, but because we built our way of working out of it.
We also teach: strategic advisory for more established organisations and hands-on training for teams who want to build this way themselves. It runs alongside these three, not as one of them.
The tools already exist. The challenge is deciding with a method:
Which processes gain something concrete, and which ones are better left for later.
Whether an assistant is enough, an automation is needed, or it takes custom software.
Which data stays protected inside the company, and who verifies the output.
Hours saved, steadier quality, margin recovered: figures that can be checked.
Who uses it every day, who supervises it, who answers for it.
What to do now, what to postpone, and what to leave out of the path.
Two recurring challenges: recognising which processes are worth redesigning before automating them, and telling apart the return promises that can actually be demonstrated.
Every activity finds its place on the map by answering two simple questions: how clear is it, and how repetitive is it. The position it lands in determines the right kind of intervention.
Redesign the flow first, then consider automating it.
This is where we start: automation, workflow or a dedicated assistant.
It deserves a full redesign before any technology touches it.
Drafts, checklists and decision support. The person stays in charge.
When the process is sound first, automation multiplies the results.
Each phase has a contained cost and a short timeline, and is built from the start as part of a single system. Each phase unlocks the next: you move on based on the results already achieved.
Understand which processes stand to gain something real from AI.
The map of where the gains are, the top three priority use cases, the roadmap and a realistic cost range.
Get visible benefits without touching any existing IT system.
Written procedures, ready-to-use templates, hands-on team training and the first measured savings.
Automate one or more recurring flows, only on processes already validated.
A working, documented and monitored automation, with far less recurring manual work.
Build software only when volume and value justify it.
A system the company owns: interface, data, integrations and specialised assistants.
Everything we decide together — in a meeting, a call or a review — becomes a requirement, and every requirement gets built. Nothing gets lost between the conversation and the software.
What we agree on becomes a user story, a page in a shared knowledge base — not a note nobody finds again.
Every new request starts from what's already there. It's why the second month costs less than the first.
What gets described gets shipped — including the agents that live inside the software, when a system works better with one.
The steps that matter always wait for someone to approve them. That part we're happy to keep slow.
The method is ours, and we use it every day — which is why we can hand it over.
For more established organisations bringing AI into their own processes, with someone who's done it before.
For teams who want to work this way: from using AI well day to day, to governing it while building software.
Some clients want us to build. Some want to learn to build. Both are fine.
Innovation works best when it is visible. We hand over a glass box: every result arrives with the reasoning that produced it.
You understand where we intervene and why it pays off, explained in plain language.
The numbers come from a traceable calculation that returns the same result from the same data.
The decisive steps always wait for a person to approve them.
Procedures and codified knowledge stay an asset of the company, usable well after we're gone.
A protection model built for smaller companies: the limits on AI in the business are set at the outset, in writing.
Critical and strategic data lives only on approved company tools, under the company's control.
AI proposes drafts and options. Whoever approves, signs and takes responsibility is always a human being.
Full isolation between everyone's personal apps and the governed company systems.
Every AI-assisted process has a named owner and a result to keep an eye on.
We start by reading the processes. Technology comes afterwards, on the points the analysis has already flagged.
A first conversation about your business context, the real priorities and the goals. No commitment.
A structured interview on your processes: what works, what jams, and how ready the company is.
The operating plan: priority use cases, reference metrics and the resources it takes.
Three quick details. We'll get back to you within one business day.