2026 / Lead Product Designer
AI-Driven Maintenance Assistant
Natural language over maintenance data — an AI concept taken from idea to live trade-fair demo in 24 days, opening the sales conversations the platform needed.
Context
Maintenance platforms hold the answers — behind filters, forms and reports. The people running rail operations ask questions in words: which vehicles are overdue, what failed last month, where is the risk. The business needed to show, credibly and soon, what an AI-native answer to that gap looks like — with a hard deadline: the industry trade fair.
Highlights
Key decision
Scope to one honest demo. Real questions, real maintenance data, answers that cite their records — and nothing staged. An AI concept a rail operator can interrogate at a fair booth is worth more than a roadmap slide, and it cannot be faked.
Process
Twenty-four days from idea to a working demo on the fair floor. The conversational flows were mapped to the platform's own nouns and verbs — assets, defects, work orders — and iterated daily with domain experts, so the assistant answered like the industry speaks.
Results
24 days
from idea to live trade-fair demo
The assistant anchored the trade-fair presence and turned booth visits into concrete customer conversations about the AI roadmap — the outcome the concept was built to produce.
Credits
Concept & design — Uroš Šolar · built with Boom Software R&D