Understanding Reference Material and AI Workflows in Lorikeet
Gain insights into how Lorikeet leverages reference material and AI workflows to enhance customer service responses.
This article provides an in-depth look at how Lorikeet uses reference material and AI workflows to enhance customer service responses.
FAQ Responses in Lorikeet
Lorikeet uses a technique for its FAQ responses that leverages a database of reference material. Here’s how it works:
- A corpus of reference material (e.g. help center articles) is collected and stored
- When a user asks a question, the AI searches this database for relevant information
- The AI finds an answer based on the similarity in meanings between the question and the stored content
- Both the original question and the retrieved information are then used to generate a response
Advantages of FAQ Responses:
- Provides up-to-date information from your knowledge base
- Works well for simple, straightforward questions
- Can leverage existing documentation quickly, with minimal setup
Limitations of FAQ Responses:
- Not ideal for issues requiring multiple steps
- Not capable of looking up specific customer information or taking actions
FAQ Responses vs. AI Workflows in Lorikeet
While FAQ responses are powerful, they’re just one tool in Lorikeet’s arsenal for answering customer queries. Let’s compare FAQ responses to AI workflows:
When to use FAQ Responses:
- One-shot, single-turn conversations
- Simple questions with consistent answers across users
- Examples:
- “What are your opening hours?”
- “Can you give me a call?”
- “How much does this service cost?”
When to use AI Workflows:
- Complex queries requiring multi-step logic
- Questions that depend on user-specific data or context
- Scenarios involving multiple data sources or APIs
- Example:
- “Where is my order?” (requires access to order history, delivery times, etc.)
In Lorikeet, we typically attempt to match queries with specific workflows first. If no suitable workflow is found, the system falls back to the FAQ responses, ensuring that users always receive the most appropriate and accurate response possible.
| Feature | FAQ Responses | AI Workflows |
|---|---|---|
| Best for | Straightforward questions | Complex queries with multiple steps |
| Data sources | Pre-existing reference material | Multiple data sources, APIs, user-specific data |
| Flexibility | Limited to available reference content | Highly flexible, can incorporate custom logic |
| Update frequency | Can be updated frequently (e.g., every 15 minutes) | Requires manual updates to logic and configuration |
| Customization | Based on provided reference material | Fully customizable logic and responses |
| Fallback option | Often used as a fallback when no workflow matches | Primary option for all queries |
By utilizing both approaches, you can create a robust and efficient customer service experience that adapts to a wide range of scenarios.
Was this page helpful?