Alongside systems built for conventional customer care tasks, FPT.AI also provide a more comprehensive customer care solution using AI, which helps reduce workload for human staff, at the same time improve work efficiency for the business.

Customer care is a critical aspect for all businesses in all fields. In systems for customer care, customers are often connected to call agents, who will record their arisen problems and assist them to solve these problems in a quick and effective manner.

AI virtual agents, meanwhile, are gaining popularity, and are expected to bring fresh customer experiences. These agents are also able to evaluate customers with lead qualification, collect data, confirm orders, and even solve problems – all lead to attractive experiences with good results, while saving both cost and time for businesses.

In reality however, customers often make complicated queries that chatbot cannot process. This is why FPT.AI provide another tool that can redirect these queries from chatbot to human staff, for the fastest and most immediate request recording and assistance.

Below are how to customize the chatbot on the FPT.AI system:

1. Establish scenarios for triggering connection to human staff

For scenarios that trigger this connect, FPT.AI chatbot issues the “Request for human support” card to notify the system. The card can be placed in any scenario or location that requires support from human agents, and with it, you can choose to stop the bot from answering customers, replacing them with call agents.

2. Customize working hours

To customize working time of the bot, FPT.AI system already includes the following time variables, with time set at UTC+0:

Variable Description Example
 

sys_datetime_date

 

Date in y/m/d formate  

2020/08/04

 

 

sys_datetime_date_year

 

Year  

2020

 

 

sys_datetime_date_month

 

Month  

8

 

 

sys_datetime_date_day

 

Day  

4

 

 

sys_datetime_date_year_day

 

Date in year from 1/1  

217

 

 

sys_datetime_date_week_day

 

Date in week with Sunday = 0  

2

 

 

sys_datetime_hour

 

Hour  

15

 

 

sys_datetime_minute

 

Minute  

42

 


For utilization, you can declare one of the above variables. Upon declaration, you will be able to use these for condition review for enforcement of cards in a scenario. Please notice that these variables are all at UTC+0, while Vietnam is at UTC+7

For convenient customization of conditions, declare the variable is_working_time, which will record if the current time is during the working hours.

minute_of_day = sys_datetime_hour*60 + sys_datetime_minute

In some cases, working hours may include minutes, e.g. 7h30. Configuration for such conditions is quite complex, and therefore I will make it about calculating working time in a day by calculating the minute of day at (UTC+0). We will also declare a variable like minute_of_day for use in future scenarios.

Let’s set the chatbot working hours to be from 7h30 – 20h in Vietnamese time, which is 0h30 – 13h at UTC+0, and by minute, that is from the 30th minute to the 780th minute of the day.For example, working hours started at 7h30 which is at the 7*60 + 30 = 450th minute. This simplification will make it much easier to set complex working hours.

The is_working_time variable will be customized like in the below image:

Next, we will customize a scenario for connecting customers with call agents during working hours, and notifying them to contact other support channels if out of work hours like below:

And we are done! Conditions in reality maybe a tad more complicated, as we will need to check for holidays, breaks, and so on. This can be done using variables like date, month, date in year, date in week…

3. Conclusion

The combination of chatbot and human call agents, as well as customization of appropriate time in support scenarios will help businesses in reducing and effective using human call agents, at the same time improving customer experience.

In operation, not all businesses have 24/7 ready staff for customer support, but rather only during particular hours. The FPT.AI chatbot is able to solve such a problem, providing in-time effect support for customer,s with minimal cost.

Other than artificial intelligence, FPT.AI chatbot is also integrated with state-of-the-art natural language processing (NLP) technologies. The bot interface is also regarded as easy to use with compatible visuals. With the system, users can easily create bot scenarios with just simple actions.

Easily integrable to businesses’ systems and popular chat interfaces, FPT.AI chatbot has become partners of many large companies and corporations, help to free human resources, optimize work procedures, thus generating huge and robust values through time.

The previous article had discussed how to integrate AI chatbot into business systems and enable the bot to automatically issue voucher codes for customers. The full article can be found here.

Bach Pham – FPT HO

Related posts: