ClouderaNOW Activate your data, unleash your AI, and scale across any cloud | Dec 9

Register
  • Cloudera Cloudera
    • Blueprint Card

    Overview

    Deliver personalized financial product recommendations with AI-powered customer engagement, built-in risk controls, and end-to-end visibility on Cloudera.

    Financial institutions are under increasing pressure to deliver personalized customer experiences while maintaining strict controls over product eligibility, financial risk, and regulatory requirements. Traditional customer engagement channels often struggle to provide timely, relevant recommendations at scale, while conventional AI chatbots can lack the transparency and safeguards required for sensitive financial interactions. 

    The Agentic Next-Best-Action with LangSmith Observability Blueprint combines Cloudera AI with intelligent conversational agents and LangSmith monitoring to deliver personalized credit card recommendations based on customer needs and predefined eligibility criteria. By separating AI-generated conversations from business-controlled decisions, organizations can improve customer engagement, accelerate product recommendations, and maintain visibility into how agents operate. Built-in evaluation and monitoring capabilities help teams continuously assess agent performance, identify issues, and improve the reliability of AI-powered customer interactions.

    Business Value:

    Accelerate personalized financial product recommendations with agentic AI on Cloudera and LangSmith Observability, increasing cross-selling opportunities, reducing customer engagement costs, and improving AI accountability through built-in risk controls, transparent decision-making, and continuous performance evaluation.

    • Cost Savings - Reduce customer service and sales costs through automated product recommendations and conversational engagement; Improve relationship manager productivity by automating routine customer qualification and product inquiries; Lower AI troubleshooting and monitoring costs through centralized agent observability and performance tracking; Accelerate AI development and testing through reusable evaluation workflows and automated quality assessments.

    • Cost Avoidance - Reduce risks of unsuitable financial product recommendations through predefined eligibility rules and customer risk screening; Minimize compliance and operational risks through traceable AI interactions and auditable decision workflows; Prevent costly AI deployment errors by evaluating recommendation accuracy and guardrail effectiveness before introducing changes.

    • Revenue Growth - Increase credit card cross-selling and upselling opportunities through personalized, eligibility-aware recommendations; Improve conversion potential by aligning financial products with individual customer needs and preferences; Scale personalized financial product distribution without proportional increases in sales and customer service staffing.

    Demo

    This demo shows how financial institutions can use Agentic Next-Best-Action on Cloudera AI with LangSmith Observability to deliver personalized credit card recommendations while maintaining control over customer eligibility and AI performance.

    See how an AI-powered assistant engages customers in natural conversations, asks questions to understand their financial goals, and recommends appropriate credit card products based on predefined business rules. Customers who do not meet established risk criteria are safely redirected to a human representative rather than receiving an unsuitable offer.

    The demonstration also highlights how LangSmith provides visibility into every customer interaction, allowing business and technical teams to review agent decisions, monitor response quality, capture customer feedback, and compare performance across different AI configurations. Together, these capabilities illustrate how organizations can scale personalized customer engagement while improving transparency, consistency, and operational oversight.  Watch the demo

    Key Features

    • Multi-turn LangGraph Multi-Agent System. The application processes incoming customer inquiries and proactively upsells credit card offers based on customer preferences.

    • Customers are filtered by risk and rerouted to a different marketing channel if needed. Agents interact with the customer and consult an internal rule engine to come up with the best credit card offer for upsell.

    • Agent decisions are tracked in real time via LangSmith Observability where stakeholders can monitor interactions and agent performance. Conversations tracked as root runs featuring nested runs to evaluate every piece of the interaction into a coherent logical thread.

    Use Case

    Banks and card issuers use chat to offer upgrades and cross-sells, but three requirements keep a general-purpose LLM chatbot from handling that traffic:

    1. Regulated recommendations. Every offer must be tied to a rule-based eligibility check on risk tier, age, and income, not left to the LLM’s judgment alone.

    2. Fail-safe decline path. High-risk customers must be routed to a polite decline that quotes no specific product figures and redirects them to a human agent.

    3. Auditability. Product, risk, and compliance teams need to inspect every conversation turn by turn, replay it against new prompts, and grade it offline before promoting a model change.

    This blueprint meets all three by separating the conversation from the decision: the LLM talks, a rule engine and risk guardrail decide, and LangSmith records and grades every turn. The result is a personalized offer channel that runs at chat scale while staying explainable to the teams accountable for it.

    Architecture

    The blueprint is built from five components. Four run on Cloudera AI, and LangSmith provides the observability layer:

    • Multi-agent system: the NBA engine, built as a LangGraph workflow. Each customer message passes through intake, risk screening, and intent routing, then on to a clarifying question, an offer, a follow-up answer, or a decline. Conversation state is checkpointed per thread.

    • Agent observability: LangSmith traces every turn as a root run, with each agent step nested beneath it. Turns are grouped into conversation threads, customer feedback is stored against each reply, and offline evaluation experiments run here too.

    • Model serving: Cloudera AI Inference Service hosts the Nemotron LLM behind every language step: extracting customer details, routing, asking questions, pitching offers, and judging evaluations. It can also serve the optional XGBoost or PyTorch risk classifier, exported to ONNX.

    • Data & metadata: the customer profile table that drives the risk guardrail, and the offer catalog with its eligibility rules, which feeds the rule engine. Both use SQLite in the demo.

    • Web layer: a FastAPI service that exposes the agent as a JSON API and serves a lightweight TypeScript chat UI from a single port. It’s deployed as a Cloudera AI Application with no build step.

    Get started by deploying this blueprint in your environment

    Every blueprint ships with source code, sample data, and step-by-step instructions — ready to run on your Cloudera Platform

    Your form submission has failed.

    This may have been caused by one of the following:

    • Your request timed out
    • A plugin/browser extension blocked the submission. If you have an ad blocking plugin please disable it and close this message to reload the page.