Experience

Our experience.

Projects delivered by our team in previous roles, across automotive, healthcare, events and research.

Automotive

Context: In an automotive company, engineering and management worked on two separate systems: Teamcenter for projects, SAP for management.
Problem: Information had to be moved by hand from one system to the other, with a lot of coordination work.
Solution: An application that connects the two systems and transfers information automatically.
Result: Project progress, bills of materials and reports always aligned, in one place.

Healthcare

Context: Certified medical software has to be tested by simulating what a user does. These tests were written by hand.
Problem: Writing and checking tests by hand was slow, and there was no way to know whether they covered every case.
Solution: An application that reads the code and a sample test, understands how it works and writes new ones on its own.
Result: Tests generated automatically, bugs found, missing cases checked and new tests proposed.

Events

Context: An event with many areas and roles: administrators, organisers, area leads and people who only need to look things up.
Problem: Schedule, activities and materials were scattered across areas, and everyone had to edit only their own part.
Solution: An app with each area's schedule, a voice assistant to add activities by speaking, and inventory management.
Result: A single dashboard: everyone sees what they need and edits only their own area.

Healthcare

Context: Emergency department records, coded internally since 2008, had to be converted to the international OMOP standard.
Problem: Italian terms had to be matched to more than 2 million entries: traditional software finds the right one in only 67% of cases.
Solution: A group of AI agents that translates the terms and finds the matching entry in the standard.
Result: The right entry is among the top 5 suggestions in about 97% of cases, versus 67% for traditional software.

Research

Context: Ongoing research project: library documents, some in Arabic, that an expert had to catalogue and check by hand.
Problem: Manual cataloguing takes a great deal of time, and automatic text recognition (OCR) introduces errors.
Solution: Several AI agents recognise the work even from poorly read text and retrieve correct data from reliable catalogues.
Result: Records already filled in with verified data: less work for the expert and fewer errors.

Research

Context: Research track: getting data out of a database by asking questions in everyday words.
Problem: Getting data out of a database means writing SQL. Example: “Which Formula 1 driver won in 1998?”
Solution: The AI turns the question into SQL, runs it and returns the answer.
Result: Anyone can get the data with a simple question, without writing code.

We build solutions
that create value.

Got a knot to untie? Tell us in a few lines.

We reply within two working days