Interactive Flow
Interactive Flow includes a visible chat interface, allowing users to interact with the AI workflow directly through a chat window.
Key concept for Interactive Flow
- Auto Continue: The workflow can continue executing in the background without user input.
- Human in the Loop: Humans can make the final decisions in complex scenarios where AI models are uncertain.
Use Cases
Online Customer Service: Users can ask questions in the chat window, and the Interactive Flow can retrieve relevant data from the internal knowledge base to provide detailed responses.
Internal Help Desk: The LLM analyzes the context and intent of the employee's query in real-time. Employees can receive instant and accurate response to common questions, such as IT support, HR policies, etc.
Service Flow
Service Flow operates without a visible chat interface. It typically functions as a background service or an API endpoint, triggered by various events.
Key concept for Interactive Flow
- Trigger/Timer: Initiate a workflow on a schedule, or trigger by webhook event, message events.
- Workflow as a Tool: The workflow can be published as a tool, and it can be used in other workflows.
- API Call: The workflow sends a request to an external API, specifying the needed data or action.
Use Cases
Sales Email Automation This Service Flow use LLMs to automate writing sales emails, just provide details like the recipient's name, company and key information, it can generate personalized email content for your customers. Users also can set a daily schedule to automate follow-ups.
Market Research Report Generation: The Service Flow gather data from various sources such as surveys, social media, sales databases. Use LLM to identify trends, patterns, and key insights from the data, then it can automatically generate a comprehensive market research report.
Differences Between Interactive Flow and Service Flow
Key Differences | Interactive Flow | Service Flow |
---|---|---|
Interface | Visible chat interface for user interaction. | Interface-free operations |
Trigger | Triggered by user input in the chat interface. | Triggered by events such as scheduled tasks, Webhook events, or external system calls. |
Application | Suits for real-time interactive scenarios. | Ideal for automated operations and backend services. |
Benefits | Enhance user experience | Optimize operational efficiency |