[mercari GEARS 2025] Transforming customer engagement with Google Customer Engagement Suite

Transcript

1. Peranikov
 Mercari / Software Engineer, Contact Center
 Transforming Customer Engagement with Google Customer Engagement Suite

2. Joined Mercari in 2018 and has been primarily working as a Backend Engineer, developing services related to the customer service domain. In 2024, transferred to the Contact Center team that develops systems for handling customer inquiries, and currently serves as the Tech Lead for projects that transform the customer service experience.
 @Peranikov
 
 
 Software Engineer, Contact Center

3. 1) CS Engineering Initiatives
 2) Customer Support Challenges and Our Improvement Project
 3) System Architecture and Walkthrough
 4) Closing
 Today’s agenda

4. CS Engineering Initiatives

5. Our CS Engineering division develops the following systems to resolve issues for Mercari users:
 • The Help Center to resolve customer problems
 • A system to handle incoming customer inquiries
 • A tool to manage and view customer information
 
 CS Engineering Initiatives

6. Help Center

7. Customer Support Challenges and Our Improvement Project
 

8. Customer Support Challenges
 
 “We are slow to resolve issues…”
 
 • Self-resolution through FAQs is burdensome for customers
 • Email-based support can take up to 50 hours to resolve a single issue
 
 

9. “Effortless Customer Experience Project”
 
 And we defined the goal …
 “Solving any Mercari customer's problem within five minutes, anywhere in the world.“
 Our Improvement Project
 

10. Transition to a Chat-Based Inquiry System
 🔍 Easily identify problems
 Instead of searching for FAQs, Mercari users can use a chat-like experience to identify the problem.
 󰢧 Enable personalization
 The system retrieves customer data from various microservices, analyzes the situation, and resolves issues on the spot without needing an operator.
 ⏱ Provide real-time solutions
 Complex problems that cannot be resolved by the system alone are quickly solved in real-time with operator assistance.

11. Adopt Google’s Customer Engagement Suite 
 
 Source: https://cloud.google.com/blog/ja/products/ai-machine-learning/introducing-customer-engagement-suite-with-google-ai
12. Why adopt it?
 • Seamless integration with our existing GCP-based microservices architecture.
 • Easy integration with other AI solutions like Conversational Agents.
 • Dedicated support from the Professional Services Organization (PSO) to reduce the risk of adopting a new product.
 
 Adopt Google’s Customer Engagement Suite 
 

13. System Architecture and Walkthrough

14. Product Demonstration - Chat Support

15. Product Demonstration - Chat Support

16. Product Demonstration - Chat Support

17. Product Demonstration - Chat Support

18. Product Demonstration - Live Contact Tool
 

19. Product Demonstration - Live Contact Tool
 デニムパンツ
20. Product Demonstration - Live Contact Tool
 デニムパンツ
21. Product Demonstration - Live Contact Tool
 デニムパンツ
22. Product Demonstration - Live Contact Tool
 デニムパンツ
23. System Architecture (Chat with Conversational Agents)
 Help Center Frontend Customer Contact Kubernetes Cloud Storage Cloud Spanner micro services Kubernetes Headless Web SDK Cloud Pub/Sub CCaaS Conversational Agents BigQuery External Storage
24. Headless Web SDK
 
 • Headless Web SDK is an SDK for
 handling conversational operations.
 • CCaaS also offers a Web SDK that
 includes a UI
 • We chose to use the Headless because
 we wanted to incorporate our own custom
 UI components.
 

25. Conversational Agents
 
 • Chat bot uses the rule-based Conversational Agents.
 • Flow development is handled by non-engineers.
 

26. External Storage
 • CCaaS includes a feature for exporting chat conversation data and associated metadata to Cloud Storage.
 • With the "Raw data export" functionality, CCaaS can integrate statistical information such as Handling Time.
 • Our product utilizes these features to allow operators to access historical inquiry information and enable the metrics team to conduct performance analysis.
 

27. Help Center Frontend System Architecture (Chat with Human Agents)
 Live Contact Tool Frontend Headless Web SDK Customer Contact Kubernetes Cloud Spanner Operator Agent Adapter Others Kubernetes CCaaS
28. Agent Adapter
 • The Agent Adapter can
 be embedded in the
 front-end using an
 iframe.
 • The front-end fetches
 data and switches
 panels, triggered by
 events from the iframe.
 
 デニムパンツ
29. Closing

30. • Customer issues were taking too long to resolve.
 • We defined the goal “Solving any Mercari customer's problem in five minutes, anywhere in the world.“
 • We transitioned to a chat-based system to create a foundation for AI-driven problem resolution.
 Closing

31. • Integrate AI solutions provided by Customer Engagement Suite into our products.
 • Expand areas where customers can resolve issues through self-service.
 • Provide multilingual support for customer inquiries.
 Future Outlook…

32. Thank You!