Chatbots vary widely depending on the use case and complexity level required when deploying a help resource for users.
They can be anything from simple chatbots which are driven by keywords to much more complex AI-powered chatbots which are protected by several layers of authentication and integrated with external systems.
Keyword-driven chatbots work on keyword recognition. These are the words that you are adding to the chatbot's backend configuration and providing synonyms of those words to be treated the same as the main keyword.
This is a much simpler version of a chatbot, but must additionally cater to as many user responses as possible to avoid ending up on a fallback loop.
Intent-driven chatbots are driven by the Infobip AI engine and work on a combination of keywords and machine learning to achieve smart, and almost self-service, chatbots.
There are several ways you can protect your chatbots and the content they capture for use cases which are likely to be dealing with more sensitive information.
Use any or all of the secured chatbot options to add extra layers of security to specific attributes, dialogs, or even entire chatbots.
Authenticated dialogs allow you to secure specific parts of the chat with end users for more sensitive parts of the conversation for example when discussing personal details.
Conversations which are protected with authenticated dialogs require users to authenticate themselves using criteria that allow them to identify themselves when checked against records in an external database.
It works by redirecting users to an authentication window when landing on a dialog which is secured by authentication. If the background _authentication_ attribute is set to FALSE, the user will be required to authenticate themselves. If set to TRUE, i.e. they have already authenticated themselves, they will be able to continue the conversation.
Authentication can be set up to expire are a set amount of time so that should users let the authentication lime lapse (e.g., they leave the chat open), they will be required to re-authenticate.
Default and Expiration dialogs can not be restricted with authentication. Use the To Dialog element to route users to authenticated dialogs.
Create a chatbot with authenticated dialogs
Use Apple's exclusive Apple authentication when using Apple Messages for Business as your chatbot channel. This allows you to authenticate users through external databases all within the Apple chat.
Authentication through API
Use the API element and configure authorization headers and response settings to GET user information from your own databases, or 3rd party databases.
When using attributes to store potentially sensitive data, you can secure them so that whenever they leave the database values are encrypted, not visible to the human eye.
They work exactly the same as all other attributes but have a setting enabled which will tell the system that these attributes are sensitive. This is especially useful when chats are being transferred over to live agents.
Use the coding element to introduce your own means of security around chat content and chatbots.
Create keyword-driven chatbots in any language you require and deploy it to any market in the world.
Intent-driven chatbots currently support a limited number of languages but continue to be expanded.
A range of channel elements is available to enable integration to all kinds of external systems and services. Here are some supported features in chatbots:
- Code injection
- Infobip People solution
- Infobip Conversations solution
- Apple authentication
- Payment systems