Technology for People: Reflections from VESTA 2026

VESTA 2026 was held at the Auditorium della Tecnica in Rome, organized by InnovUp around the theme of care economy — the sector covering home care, elderly assistance, disability services, and caregiver coordination. I attended representing Intuitiva Salute, where I work as a developer.

It is easy to go to an industry event and come back with nothing but a stack of business cards. VESTA was different — not because of any single talk, but because of the quality of the problems being discussed. This is my attempt to make sense of what I heard, and how it connects to the work I am actually doing.

What Is the Care Economy?

The care economy is the infrastructure of human assistance. It is one of the fastest-growing sectors in Italy and Europe, driven by an aging population and shrinking public budgets. But the challenge is not just financial — it is informational.

Families trying to coordinate care for an elderly parent deal with fragmented systems: paper records, phone calls between providers, manual scheduling. The cognitive load falls on the caregiver, often a family member already stretched thin.

The interesting projects at VESTA were the ones that started from the workflow first. Not “let’s apply AI to healthcare” but “a care coordinator manages 30 cases, each with different providers, and currently tracks everything in a spreadsheet — what breaks first?” That framing is the right one. Technology does not solve the emotional weight of caregiving. But it can remove friction. That is the specific problem worth working on.

What I Am Building at Intuitiva Salute

At VESTA I represented Intuitiva Salute, a management platform for Italian medical clinics. My work spans several layers of the product, but the one most relevant to the care economy discussion is Sphera.

Sphera generates a structured medical summary of each patient’s history so the professional enters every visit with context already loaded. The problem it solves is simple: when you last saw a patient four months ago and they had three visits in between with two other specialists, you should not spend the first two minutes of the appointment reconstructing their history from scattered notes. Sphera reads the record and surfaces what matters, in a format a professional can absorb in thirty seconds.

It is not a diagnostic tool. It is a context tool. The distinction matters — especially in a sector where the line between “helpful summary” and “clinical recommendation” carries real regulatory weight.

The architecture constraint that shapes everything else is trust. Health data is not a product feature. It is a legal and ethical obligation. Every design decision — how data is stored, who can access what, what an AI model is and is not allowed to do with a patient record — runs through that constraint first.

This means no cloud AI touching patient records without explicit consent and proper data processing agreements. It means audit trails on every sensitive operation. It means role-based access that mirrors real clinical hierarchies, not simplified user-admin binaries.

The interesting engineering problems in health-tech are not about scale in the Twitter sense. They are about correctness, traceability, and the cost of a bug — which in healthcare is not a degraded user experience, it is a missed appointment or a wrong record pulled at the wrong moment.

AI in the Care Economy: Social by Design

One of the recurring themes at VESTA was AI as an enabler for care workers, not a replacement. The most compelling use cases were mundane: automatic summarization of patient histories before a visit, smart scheduling that accounts for travel time and care continuity, anomaly detection in vitals for home monitoring.

These are not flashy demos. They are tools that give care workers back minutes in a day — which, at scale, translates into better care and lower burnout.

The principle I keep returning to: AI in this sector has to be social by design. The model output is only as valuable as the trust the care worker places in it, and trust is earned through transparency, not through accuracy benchmarks alone. A system that is right 95% of the time but opaque about the 5% is actively dangerous in a clinical context.

The hardest part of building AI for people is not the model — it is the interface between the model’s confidence and the human’s judgment.

What VESTA Was Good For

Beyond the talks, the most valuable part was exposure to teams working on adjacent problems: home care logistics, caregiver matching platforms, remote monitoring infrastructure. None of it was directly applicable to what I am building today, but some of it will be.

Health-tech in Italy is fragmented enough that the sector genuinely benefits from projects finding each other. The collaborations that will matter in a few years are probably seeded at events like this one — not through formal partnerships announced on stage, but through conversations about specific technical problems.

Three Things I Came Back With

The system includes the people. A booking platform is not just a database and an API. It is a tool that a receptionist uses under time pressure, that a patient navigates when they are anxious, that a doctor checks between appointments. Designing without those people in mind produces technically correct software that fails in practice.

Governance is architecture. Who owns the data, who can see it, how long it lives — these are not compliance checkboxes. They are load-bearing decisions that shape every layer of the stack above them. Getting them wrong early is expensive to fix later.

Friction is a design choice. Every step removed from a care coordination workflow is a decision worth defending. Not all friction is bad — some is protective. The discipline is knowing which is which, and being honest about the tradeoff when you remove it.

The future of healthcare in Italy is digital. Whether it will also be human depends on who is building it, and what constraints they take seriously from the start.