Demo: Build Secure Open Banking Consent Workflows with Agentic AI and MongoDB Atlas

GitHub: - Backend: https://github.com/mongodb-industry-solutions/open-finance-next-gen - Chatbot: https://github.com/mongodb-industry-solutions/leafy-bank-backend-openfinance-reactagent-chatbot - UI: https://github.com/mongodb-industry-solutions/leafy-bank-bian Sign-up for a free cluster → https://www.mongodb.com/cloud/atlas/register Subscribe to MongoDB YouTube→ https://mdb.link/subscribe Open banking brings financial data together—but customers should understand and control what an AI agent can access. This walkthrough shows a consent-first, multi-agent workflow for explaining data scope, purpose, institution, and access duration; collecting explicit approval; pausing for bank authentication; and revoking access when it is no longer needed. Built with MongoDB Atlas, LangGraph, MongoDB Queryable Encryption, and the MongoDB MCP Server, the demo combines persistent workflow state, protection for sensitive fields, and agent access to a limited set of read-only database tools. The example uses fictional Leafy Bank and open finance; the underlying pattern can be adapted to other applications that need transparent approvals and controlled data access. Watch the architecture walkthrough and live demo to see the consent journey end to end—from choosing what to share to querying consented data and revoking access. Solution Library: https://www.mongodb.com/docs/atlas/architecture/current/solutions-library/open-banking/ Blog: https://www.mongodb.com/company/blog/innovation/powering-open-banking-consent-with-agentic-ai White papers: - Next Generation Open Finance with MongoDB and Agentic AI: https://www.mongodb.com/resources/solutions/use-cases/next-generation-open-finance-mongodb-agentic-ai - How Latin American Countries are Shaping the Open Finance Revolution: https://www.mongodb.com/resources/solutions/use-cases/latin-america-open-finance Chapters 00:00 Why AI data access needs consent 01:04 The open banking consent experience 02:12 Making consent clear, granular, and revocable 02:29 A conversational consent copilot 03:21 Multi-agent workflow architecture 03:53 MongoDB Atlas as the data layer 05:08 Queryable Encryption for sensitive consent data 05:56 Read-only AI data access with MongoDB MCP 06:30 Live demo: customer consent and bank connection 07:48 Querying data and revoking consent 08:12 Key takeaways: controlled, auditable access 09:02 Architecture and code repositories Visit Mongodb.com → https://mdb.link/MongoDB Read the MongoDB Blog → https://mdb.link/Blog Read the Developer Blog → https://mdb.link/developerblog MongoDB for Developers YouTube Channel → https://www.youtube.com/@MongoDBDevelopers