1. Applications of chatbots within enterprise environments

Many enterprises and businesses in the fields of finance and banking, insurance, and retail are currently utilizing chatbots as they effectively replace repetitive tasks, help saves time and human resources cost, as well as further optimize the businesses’ productivity.

In the Service – Retail sector, chatbots can help customers: Look up product information; Check warehouse status; Look up the stores’ addresses and make suggestions for the nearest store based on the given information; Order items online, Book and cancel airplane tickets, Search for flights information and booking codes, as well as information regarding promotions campaigns, etc. 

Within the field of Finance, Banking and Insurance, chatbots are utilized for the following purposes: Swift and timely customer support 24/7;  Assist customers in looking up for interest rates, service packages, loan programs, or open credit cards, etc. Regarding customer support, chatbots play an essential role as they help improve user experience, and therefore, help the enterprise build an enthusiastic and modern image. 

The chatbots will automatically store the information provided by users in the CRM system for analyzing and effective data management purposes. This is a highly useful feature that can assist enterprises in developing more customer-friendly chatbots, as well as building more customer-focused marketing strategies.

The FPT.AI chatbot, equipped with artificial intelligence, big data, and machine learning technology, has self-learning ability that allows them to become smarter after each conversation with customers.

2. Easy creation of chatbot with FPT.AI Conversation

FPT.AI Conversation provides a chatbot creation and management platform with a user interface. By implementing natural language processing technology and dialogue management system, the platform provides a complete chatbot building solution that can be integrated across various platforms and chat interfaces. These technologies also allow the platform to assist the automation process of businesses and customers’ conversations, contribute to sales growth, expanding the business’ scale of operations, and improve customer satisfaction.

APIs (Application Programming Interface) are methods and protocols used to connect libraries and different applications. APIs allow users to gain access to a set of commonly used functions, from which data can be exchanged between applications. To gain access to FPT.AI’s APIs, you need to create an account on bot.fpt.ai. If you don’t have an FPT.AI account yet, create one now to proceed to the next steps.

3. Guide to chatbot integration with API connections

In the actual construction of chatbot scenarios, in addition to the chatbot scenarios that answers Frequently Asked Questions where the answer is already defined in the script, some chatbot scenarios require access to the enterprise database in order to complete the answer sent to customers. For example, once a customer has successfully registered their personal information, the chatbot needs to record these data and push the data back to the enterprise’s management system. After that, the CRM Card feature will help businesses perform the next tasks for customers.

With FPT.AI Conversation, the scenario builder can take an active role in designing the flow to determine the time and the data needed to interact with the system through JSON-API tags.

Below is the technical model FPT.AI Conversation is currently implementing:

In this model, in order to connect to the enterprise’s backend system, we need to build a component called middleware, which takes on the tasks of:

  • Act as an intermediary component for the chatbot to communicate with the enterprise backend system, perform the protocol conversion, if needed, for example from SOAP, to RESTful.
  • Act as a component that provides APIs that follow RESTful standard, and comply JSON format specified by the FPT.AI platform in order to integrate into chatbot scenarios.
  • Act as a component that connects the enterprise backend system to the chatbot to perform business-related tasks that are defined on the chatbot script.
  • Carry out additional functions such as statistics, creating reports, etc.

This middleware component will be deployed within the enterprise environment.

For simplicity, in this article, the author will solve the problem of providing the customer with a voucher code once they have completed a scenario (assuming the scenario is clicking the voucher button).

Firstly, we need to define the scenario and declare the variables:

When customers click on the “Get Voucher” button, the chatbot will carry out the “Get voucher” step as follows:

In this scenario, we can flexibly pass more variables to the API as needed, be it system-defined variables or variables defined by the chatbot builders. We can also add return variables from the API to handle the scenario more flexibly and further tailored the answer to fit the needs of the enterprises.

JSON-API tags allow you to use variables or constants in the URL and headers; for example, https://api-url/{{sender_id}}. To minimize the need to change these values when working in different development environments (dev, prod), you should declare constants and use them for this case as the API-KEY example in the image above.

With the HTTP POST method, please be mindful that the system will use content-type=application/json. The returned results by the APIs will follow the following structure:

The bolded attributes require proper naming. The set_attributes attribute is a list of variables to return to the bot, and the list follows the following structure:

APIs also allow returning direct message to customers; the list of message follow the above structure. If the message includes buttons and payloads, the payload will be a string with this structure: <Step name> # base64 (payload_data), where payload_data is a JSON string with a list of variables similar to the set_attributes structure above. When the customer clicks on the button containing the payload data, the value of the variable will be passed to that step and can be used as a standard variable.

For example, we can implement the above API that retrieves the voucher code through the POST method in PHP as follows: 

Just like so, we have integrated the chatbot into the enterprise backend system. You are free to build your APIs to help the chatbot perform more specialized tasks.

Once completed, the user interface will display as follows:

Chatbots built with FPT.AI Conversation can be easily integrated with popular chat interfaces such as Facebook Messenger, Zalo, Website. This ability allows users to interact directly with businesses through natural dialogues, offer one-on-one support, and is its way to become a competent “virtual assistant” that can bring customers useful experiences.

Bach Pham – Thao My

Related posts: