In this tutorial, we will create a simple conversation bot and integrated to NL studio. For pre-condition, you must create a bot by follow this tutorial in here Bot studio tutorial. You will create a NLModel which able to process pizza ordering orders. Further explanation for NLStudio, you may see in here NL Studio Guideline
Let’s start !
Assumed you already logged in and visit your project. Go to NL Studio menu by click NL menu
Then, click on Create Entity to start.
Click on Create. Here is a view after you create intent
After you created the entity, the next step is data training. You have to enter a sentence to train
intent trait. Click on Training menu
Then, you are able to enter sentence to train data. Put a sentence “I want to order pizza” and press enter in your keyboard
Afterward, click on Add Trait and select
Next, click “Train” button to train entity you has created. The final display will look like this
Entity shall train a lot for precisely understanding user’s input, if you’re doing training data in several sentences.
If you’ve done enough training, you are able to try testing by click “Test NLU” button. This feature will support you to predict more data.
Then, publish your NL for save training data and NL model.
Next step is start integration with your bot using created entity. To get started, go to Bot menu and click NLUs sub-menu.
Click on Create NLU and fill in as follow
NLUid is found in NL menu in Setting. Click on NL menu, then click Settings
Afterward, copy NLUid into NLUid field in Create NLU drawer as follow
Click Create to continue. Successfully created NLUs will be appeared as follow
Next, go to Conversation Flows and click on tab Intents. Select
orderTxt that you’ve created earlier to handle user input (you may see this tutorial Bot studio tutorial), then add a new classifier as below
Click Update to save updated Intent.
Click on Publish bot to save your bot.
Voila, your bot is successfully integrated. Let’s start in bot emulator.
Contributing to the Documentation
Is something missing/incorrect? Please let us know by contacting email@example.com. If you know how to fix it straight away, don’t hesitate to create a pull request on this documentation’s GitHub repository.