How to customize a virtual agent for multiple brands

Challenge

How can I customize the output of my Cognigy chatbot to power multiple websites with unique branding, but share same flows and NLU?

Solution

In Cognigy.AI the Localization feature can be used for much more than just creating language specific versions of your virtual agents. It's also possible to use locales to power different versions of the same bot.

Let's take the use-case of having one virtual agent flow in Cognigy.AI that is powering virtual agents on two different websites. Each of these websites has unique branding, content links and even a conversation style. We can use a locale containing the content specific to each website to provide a custom output unique to that brand on each channel. 

Step 1: Create a Locale for each Brand

Select "Localization" from the resources menu in your Cognigy.AI project and create a new locale for each brand.

Visit the Localization Documentation

Step 2: Connect a Webchat endpoint for each website

Select "Endpoints" from the resources menu in your Cognigy.AI project and create an endpoint for each of the brands. If you are using the Cognigy webchat widget, select the "webchat" endpoint.

Visit the Endpoints Documentation

For each endpoint, configure the "Locale" respective of the branded locale you created in step 1.

Step 3: Localize your flow nodes for each Locale

Within your Cognigy.AI project flows, select the Locale (Brand) you wish to configure from the dropdown menu at the top of the flow editor window. 

The "Primary" locale will be indicated with a green icon. This locale is the main development locale where you can change the flow structure and add new nodes to the flow. You won't be able to do this for non-primary locales, but you can customize the content of any nodes that already exist.

Next, select the message node that you wish to localize in any non-primary locale. Click "Add Localization" and select the method to localize. Typically, "Copy from Primary Locale" will be used as a starting point. 

You can then freely customize the output of the node with the unique content for your specific brand.

Once this step has been completed, each of your webchat endpoints will deliver the brand specific content you have configured in your locales.