Natural language processing (NLP) allows a visitor to enter a message in their own words, and the chatbot will be able to understand the visitor and provide the appropriate response. Utilizing NLP will allow you to create intents and sample utterances that connect with conversation paths in your Velaro chatbots.
NLP can be used in conjunction with natural language question-answering to make use of both key intents used in your business domain, as well as frequently asked questions and their answers, as delivered by the QnA service. This integration first checks for intent, and if a match fails, then the QnA maker knowledgebase is searched and results are displayed to the visitor.
Start using NLP
To enable NLP, create a new configuration and associate it with a chatbot.
| 1 | Create new configuration | Create a new configuration for NLP that the bot will use. Once a new configuration is named the NLP configuration controls will appear. |
NLP Configuration
| 1 | NLP Configuration Dropdown | Select the NLP configuration that you want to configure. |
| 2 | Action Dropdown | Create a new configuration or remove the currently selected configuration. |
| 3 | NLP Configuration Tabs | Tabs that allow you to change your area of configuration, Intents, Entities, Prebuilt Domains, and Manage. |
Intents
| 1 | NLP Configuration Tabs | Tabs that allow you to change your area of configuration, Intents, Entities, Prebuilt Domains, and Manage. |
| 2 | Intents | Intents are outcomes that are reached using utterances that are trained in configuration. Visitors will be able to type to the bot and depending on their question or statement the bot will decide the intent by using trained utterances. You will configure utterances for each intent by clicking on the intent’s name. |
| 3 | Status | The status of an intent: if it is trained or not trained. |
| 4 | Add Intent | Add an intent with this button. When creating a new Intent make sure to give it a unique and referenceable name that identifies the customer's request. |
Utterances
Utterances are accessed by clicking on an intent as shown in the screenshot above. Once you select an intent from the intent configuration you will be taken to the utterance page for that selected intent.
| 1 | Add Utterance | Enter utterances that a visitor might type to teach the bot which Intent to select. It is suggested that you add at least 10 utterances for each intent. |
Manage NLP Configuration
| 1 | NLP Configuration Tabs | Tabs that allow you to change your area of configuration, Intents, Entities, Prebuilt Domains, and Manage. |
| 2 | Train Bot | Train the bot to teach it intents and the utterances that go with those intents. |
| 3 | Publish Bot | Publish the bot to let the bot use utterances to discern intents in a chat. |
| 4 | Test | Enter utterances to test which intent will be selected. |
Map intents to conversation paths
After your NLP Configuration is completed with intents, and utterances, and has been trained and published, you will be able to implement NLP by associating it with a chatbot. This configuration controls what action is taken after an intent is identified.
| 1 | Select NLP Configuration | Select the NLP configuration. |
| 2 | Mapping | Select Global or a specific path in the bot, to configure where an intent will be routed to. |
| 3 | Intent | Listed intents for the currently selected NLP configuration. |
| 4 | Bot | Select which bot the intent will be routed to. |
| 5 | Path | Select which path in the bot the intent will be routed to. |