Build your first automation

Let's suggest you run an e-commerce store, and many customers ask about your delivery options. Since the response is usually the same, you want to automate these requests.

To achieve this you'll need to create a new intent in the "Automations" menu and set up the action that will be triggered by this intent. Let's start with a simple text response.

Choose the "Uncategorized" folder in the intents tree and click the "New intent" button. Type in the intent name ("Delivery options" in this case) and hit enter. Now you can enter some phrases that your customers may use when inquiring about delivery options.

These phrases are called "utterances" in conversational AI, and building a list of these can be quite challenging when the amount of intents you want to automate grows. For "Delivery options" the phrases

Activechat removes the guessing from your virtual assistant development process by analyzing actual requests sent by your customers and grouping these into topics that can easily be converted to new intents. Check the "Conversation insights" section for more details.

Setting up a NLP intent definition in Activechat

We'll dive deeper into the intent definition screen in the "NLP automation" section of this manual. For now, just add some phrases from the example above, and set the Action that this intent should trigger. There are three kinds of actions that your virtual assistant can take for each specific intent:

• Send back a simple response (a phrase or a link, as in the example above)

• Trigger a specific conversation skill in your virtual assistant (more on this later)
• Trigger a live chat session with a human agent

For now, a simple response will do. Enter response text and click the "Save" button.

After a short training process, your virtual assistant will be able to respond to similar phrases automatically. Our built-in NLP (Natural Language Processing) engine will detect the right intent even if your customers use other similar phrases or mistype them. Please note that until the training is complete your intent will not be recognized yet. The training process usually takes under 5 minutes.

Example of Natural Language Understanding in the virtual assistant

Once you get familiar with Intents and Bot skills, you'll be able to build complex automations and improve your virtual assistant on an ongoing basis. More complex interaction scenarios can be built visually with our state-of-the-art chatbot builder.