Hank Times-Case Studies-AI Customer Service

Case Study

- AI Customer Service

넳 넲

Hank Times-Case Studies-AI Customer Service, J-E-Commerce

November 2022

In this AI customer service project for J-e-commerce, the customer plans to implement automatic language recognition and drive various customer service processes of J-e-commerce through the customer's voice input. Based on AI technology, AI customer service will be automatically completed based on the input of the end user and provide the end user with the feedback of the technical and business support they need.
The most important part of the customer-AI conversation model upgrade is to transform the conversation model from a question-and-answer model to a conversation-centric, process-based interaction model, which requires the design and implementation of the underlying engine that supports process execution. In order to meet the needs of supporting process-based interaction in the dialogue mode between customers and AI, the AI response interaction process needs to be supported in AI customer service. In order to meet the needs of other process execution in the system, all existing j-e-commerce standard processes need to be seamlessly connected with the AI dialogue function.

Project Information

Service Subscriptions: Customers are large e-commerce service providers, more than 580 active online business buyers, 8 million active online business buyers, and more than 500 active business sellers.

Hank Times team:
Hank Times provided the complete agile software team for this project. Including PM, solution architect, front-end developer, back-end developer, UI designer, business analyst, etc.
This is a 14-month software project.

Hank Times Service Scope:
-Co-design solutions with customers.
-Software Detail Design
-Front-end and back-end coding, QA,CI/CD, maintenance
-Business workflow design

Application Architecture

(in Archimate)

business value

online AI customer service robot
J-e-commerce online AI customer service robot is the industry's first large-scale commercial emotional robot, in retail, municipal, customer service, medical and other fields have created large-scale application cases. In all running applications, j-e-commerce online AI customer service robot has realized that more than 90% of user consultation is received by robots, reducing customer service labor costs, improving service efficiency, and providing users with 7 × 24 hours of convenient intelligent response services.
Both semantic recognition and speech recognition achieve an accuracy rate of 95%.

voice-activated customer service robot
It includes voice response (automatic transfer) and voice call-out (automatic call-out) services to provide services to users in an anthropomorphic voice interaction. At present, it has been commercialized on a large scale in e-commerce, logistics, finance and other industries.

Intelligent Workstation
We introduced AI capabilities and upgraded the workstations of the manual agents to achieve a significant increase in their efficiency.

intelligent scheduling
Optimize the scheduling of user queries and service resources, reduce user waiting time, improve user experience, and optimize the utilization of call center personnel.

intelligent quality inspection
Intelligent conversation analysis tools extract risks and business opportunities from conversation texts and recordings to help companies improve customer service quality and monitor public opinion risks.

competitive advantage
Nearly ten years of customer service experience, full-process intelligent products and equipment, high-emotion and intelligent machine service experience, and best practices for seamless human-machine integration services. The powerful technical architecture supports massive customer service concurrency, and hundreds of millions of calls per year are 100 percent secure.

function

**process engine** function overview:
1. Loading process
2. Execution process
3. Process status record
4. Return the execution result
5. Support process refresh

process engine consists of a process manager and a process executor: the process manager is responsible for loading process data and calling the process executor, and the process executor is responsible for executing the process and returning the result.