{"slug":"rabobank-embraces-conversational-banking-with-microsoft-copilot-studio","url":"https://findausecase.com/use-cases/rabobank-embraces-conversational-banking-with-microsoft-copilot-studio","title":"Rabobank Embraces Conversational Banking with Microsoft Copilot Studio","description":"Dutch bank Rabobank migrated its text- and voice-enabled customer service chatbots to Microsoft Copilot Studio, integrating Genesys Cloud for live-agent escalation, Azure AI Speech for voice, and Azure AI Foundry with Azure AI Search for compliant generative-AI fallback answers drawn from its website. The bank's three Copilot Studio agents now handle around 20,000 daily calls and 7,000 daily chats, achieving a 62% self-service rate for chats that are resolved without escalation to a live agent.","company":"Rabobank","industry":"Financial Services","country":"Netherlands","aiCapabilities":["Conversational AI","Generative AI"],"technology":["Microsoft Copilot Studio","Power Automate","Power BI","Azure AI Search","Azure AI Foundry","Azure AI Speech","Genesys Cloud","AudioCodes"],"deployment":"Public Cloud","problemStatement":"Rabobank's existing text- and voice-enabled chatbot platform, partly on-premises and vendor-maintained, was inefficient (chat and voice logic had to be updated separately even when using the same models), slow to market because it required specialized developer support, and hard to integrate with external systems and APIs needed to make conversational agents actionable.","solutionApproach":"Rabobank migrated to Microsoft Copilot Studio, building one voice-enabled agent and two text-based agents for business banking and retail customer self-service, with topics designed to solve end-to-end processes like blocking a credit card. Genesys Cloud handles initial call/chat intake and escalation to live agents via custom-built middleware bridging it with Copilot Studio; AudioCodes and Azure AI Speech-to-Text/Text-to-Speech handle voice translation. For generative AI, Rabobank built its own compliant pipeline outside Copilot Studio's built-in features: an Azure Function calls Azure OpenAI to retrieve the user inquiry from a transcript and generate an answer using Azure AI Search, which indexes rabobank.nl content scraped and cleaned daily via a custom Python pipeline, all deployed within Rabobank's own Azure subscription.","businessValue":"Since migrating to Copilot Studio, Rabobank's agents handle around 20,000 daily calls and 7,000 daily chats, with 62% of chats resolved through self-service without escalation to a live agent, and the bank can now manage and update its virtual agents without external help.","evidence":{"band":"high"},"sourceUrl":"https://learn.microsoft.com/en-us/power-platform/guidance/case-studies/rabobank-conversational-banking","dates":{"publishedAt":"2026-10-11T05:46:55.373Z","publishedAtSource":"pipeline","updatedAt":"2026-10-11T05:46:55.373Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/rabobank-embraces-conversational-banking-with-microsoft-copilot-studio. Bulk republication requires permission."}