Key results
The company
Siemens
siemens.comIndustrial automation, smart infrastructure, and energy technology provider.
Result highlights
- Elimination of no-results searches
The story
A global technology leader with 300,000 employees operating in 130 countries, managing complex product data across nearly 20 interconnected websites and ecommerce portals.
Customers struggled to navigate restrictive keyword searches, often sifting through pages of marketing content to find specific technical documentation. This friction caused users to abandon the site when they could not locate necessary information across the complex ecosystem.
The organization built an agent-based architecture using Amazon Nova models via Amazon Bedrock, orchestrated by AWS Lambda to validate, classify, and summarize user queries. A multi-agent workflow directs requests to specific knowledge bases based on user intent while guardrail agents filter sensitive or irrelevant topics. This setup replaces restrictive keyword matching with natural language processing to surface technical specifications directly.
Scope & timeline
- Search engine deployment in less than 1 year
- 11% user satisfaction improvement via AI search
Quotes
“Users told us that they often have to sift through pages of marketing content just to find technical specifications, which slows them down. We knew that there must be something better.”