The agentic moment in banking: A blueprint for better customer experiences
Summary
The banking industry is adopting agentic AI to address persistent customer experience issues, moving beyond traditional rule-based chatbots to goal-oriented, multi-step interactions. Agentic assistants understand intent, maintain memory, take initiative, and orchestrate tasks across systems, supporting complex workflows within defined policies. This shift is driven by rising customer expectations for AI-native experiences, intensifying industry competition, and advancements in secure orchestration and multi-step task execution by AI models. Microsoft proposes a three-step blueprint for adoption: starting with internal employee assistants for back-office tasks like AML routing, then deploying external customer assistants on owned channels for narrow journeys, and finally extending to third-party platforms like ChatGPT or Microsoft Copilot. Real-world examples include Commerzbank's "Ava" resolving 75% of customer conversations autonomously, ABN AMRO's "Anna" automating over half of millions of interactions, Bradesco's "BIA" achieving an 82% first-level resolution rate, and Virgin Money's "Redi" supporting millions of customers with over 90% success.
Key takeaway
For AI Product Managers evaluating customer experience solutions, agentic AI represents a significant evolution beyond traditional chatbots. Your teams should prioritize integrating agentic capabilities to deliver personalized, multi-step workflows that resolve complex customer needs end-to-end, thereby improving satisfaction and operational efficiency. Consider Microsoft's three-step blueprint to strategically roll out these solutions, starting internally before expanding externally.
Key insights
Agentic AI transforms banking customer experiences from reactive to proactive, goal-oriented interactions.
Principles
- Redesign experiences with outcomes in mind.
- Prioritize secure orchestration for agentic actions.
- AI models now support multi-step reasoning.
Method
Implement agentic AI in three steps: internal employee assistants, then external customer assistants on owned channels, and finally external customer assistants on third-party platforms.
In practice
- Use agentic AI for fraud resolution.
- Automate product discovery and onboarding.
- Streamline payments and money movement.
Topics
- Agentic AI
- Financial Services AI
- AI-powered Customer Experience
- Banking Digital Transformation
- AI Workflow Orchestration
Best for: Director of AI/ML, AI Product Manager, Executive
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by The Microsoft Cloud Blog.