0
1.1.1
Yokohama Patch 3, Yokohama, Xanadu Patch 9, Xanadu
Now Assist for Financial Services Operations (FSO) leverages advanced generative AI to tackle complex cases in the financial services industry, including credit card disputes and insurance claims. With features like agentic workflows, AI agents, disputes intake via Virtual Agent, and case summarization, you can quickly and accurately intake case data, summarize key case details, streamline workflows, and significantly reduce processing times and mistakes. These capabilities help financial institutions minimize manual effort, enhance customer satisfaction, and drive growth.
- Agentic AI for Friendly Fraud
- Assists dispute agents in handling friendly fraud transactions using agentic AI.
- Recommends appropriate actions for friendly fraud disputes.
- Enables human agents to evaluate dispute amounts, customer relationships, and the AI agent's analysis.
- Crafts professional responses to customers regarding decision outcomes in case of deflection.
- Disputes intake via Virtual Agent
- Offer an intuitive dialogue-based channel experience for customers to submit dispute details.
- Use questions required by card processing networks.
- Leverage Now LLM to rephrase these questions in a conversational format and infer answers for unanswered questions from customer responses.
- Insurance
- Parse complex insurance claims by condensing a claim record into a brief, high-level summary.
- Use tailored inputs and prompts that addresses specific lines of business for claims.
- Support for the Now Assist panel.
- Card disputes
- Parse complex card dispute cases by condensing a dispute record into a brief, high-level summary.
- Use a tailored prompt to effectively cover a wide array of dispute cases across several categories.
- Support for the Now Assist panel.
- New
- Introduced an agentic workflow "Help resolve friendly fraud disputes" and a Friendly Fraud AI agent to assist human dispute agents in handling transactions flagged as friendly fraud.
- Added a demo data KB article that the Friendly Fraud AI agent uses.
- Added a recognition word to facilitate asking questions.
- Changed
- Enhanced the utilization of dispute transactions for summarization and updated the dispute skill configuration.
- In default LLM-enabled Text Input, disabled the Enforce User Input option to allow rephrasing and asking different questions based on user input.
- Removed
- Removed the read-only restriction from sys_one_extend_definition_config and sys_generative_ai_prompt_config.
- Permissions and roles:
-
- System administrator (admin) role is needed for the installation.