Siemens Showcases 90% Autonomous Call Handling with Amazon Connect AI Agents

The production AI agent replaces traditional DTMF IVR trees with a natural language interface powered by Amazon Lex for intent detection. For common queries, the AI agent provides direct resolution via a Connect Customer knowledge base, and complex routing rules are managed through Connect Customer contact flows. This system autonomously handles 90% of inbound calls, resulting in reduced wait times and improved first-contact resolution.

A proof-of-concept AI agent automates the process of connecting callers with specific Siemens employees. It collects caller details, queries the internal employee directory in real-time via AWS Lambda, and orchestrates callbacks, logging interactions in the CRM. Key design principles include privacy guardrails and ensuring no inquiries are dropped.

This proof-of-concept aims to scale personalized outbound sales outreach by combining Amazon Connect outbound campaigns with conversational AI. The AI agent pulls customer data from the CRM, uses predictive dialing and answering machine detection, and delivers personalized dialogues to live callers.

Siemens Global Business Services, which supports over 80 countries, faced challenges with high call volumes, long wait times for simple queries, misrouted calls, and limited outbound sales capacity. They adopted Amazon Connect Customer as their AI agent solution due to its integrated conversational AI, generative AI, and telephony capabilities, noting its pay-per-use model and rapid deployment.

All three use cases share an architecture leveraging Connect Customer for telephony and contact flows, Amazon Lex for natural language understanding, and Connect Customer AI Agent for generative AI responses. AWS Lambda handles integrations with Siemens' internal systems, and Amazon Bedrock AgentCore Gateway exposes enterprise APIs as Model Context Protocol (MCP)-compatible tools.