ServiceNow AI agents cut hospital software support time by 65 percent

ServiceNow has deployed an artificial intelligence solution for a European hospital software provider that automatically resolves 72 percent of standard support tickets. The system processes inquiries in Italian, English, German, and French, reducing the average resolution time from 18 hours to 6.3 hours.

The platform, built on Customer Service Management, Now Assist, and AI Search, supports more than 1,000 hospital facilities. It consolidates over 90,000 documents into a single knowledge base, which includes software versions, configurations, clinical best practices, frequently asked questions, troubleshooting guides, and release notes. The centralized knowledge base supports 15 languages and local variants, eliminating the need for separate archives.

AI Search interprets user questions and provides contextually relevant answers specific to the hospital domain. Now Assist handles the automatic resolution of recurring tickets, while dynamic routing directs only the most complex cases to human specialists. The solution achieves a 95 percent satisfaction rate for AI-resolved tickets, compared to 88 percent for the traditional channel. For requests with a confidence match above 85 percent, the project reports no escalations.

The implementation is estimated to save 180,000 euros annually in equivalent support costs. The architecture combines technical knowledge with medical terminology and applies controls consistent with the management of sensitive information. The design is extensible to second-level support for hospitals and commercial partners, a capability already requested in France and Germany.

ServiceNow presents the model as replicable for other vertical software providers, including systems for product lifecycle management, manufacturing execution, fleet management, and financial compliance and reporting. The prerequisite for replication is a structured knowledge base and sufficiently mature recurring categories to sustain automation. The project was named a finalist for the Digital360 Awards in the AI Strategic Data Intelligence category.