How we built a custom automation application that handles thousands of Amazon Marketplace customer emails every month — completely unattended.
The Project at a Glance
| Metric / Attribute | Operational Detail |
|---|---|
| Daily email volume | Approximately 150 – 300 |
| Monthly email volume | Approximately 4,500 – 9,000 |
| Processing | Fully automated |
| Human intervention | Not required for routine processing |
| Response management | Database-driven |
| Operation | Unattended background service |
1. The Challenge: A Repetitive Task Consuming Valuable Time
Our client, a retailer selling products through Amazon Marketplace, regularly received customer inquiries and messages through the marketplace’s email system. These messages covered a variety of topics, including product questions, order issues, and customer service concerns.
Before our involvement, every incoming email required an employee to review the message, understand the inquiry, determine an appropriate response, and send it manually. With approximately 150 to 300 emails arriving every day, this created a recurring administrative burden. Although responding to an individual email might take only a few minutes, the cumulative effort represented a significant amount of time that employees could otherwise spend on more productive business activities.
There was also another consideration: Amazon requires sellers to respond to buyer messages within 24 hours. Failing to meet this threshold drops your response rate below the 95% benchmark and can impact your Account Health. Delayed or unanswered messages could negatively affect the client’s seller performance. The client needed a solution that would eliminate the repetitive manual work, handle incoming inquiries consistently, and operate without requiring someone to continually monitor or initiate the process.
2. The Solution: A Custom-Built Email Automation System
Rather than introducing a generic third-party automation platform, we developed a dedicated application specifically for the client’s business requirements. Written in Go, the application operates as a background service, continuously handling incoming customer emails without requiring employee intervention.
The application automatically:
• Retrieves incoming customer emails through a mail pipe.
• Examines email subject lines to identify the nature of each inquiry.
• Assigns messages to predefined response categories.
• Selects an appropriate response from a library of prepared messages.
• Sends the response through a dedicated Amazon email account.
• Copies every outgoing response to the client’s dedicated monitoring email address.
The entire process happens automatically. Employees no longer need to open each email, determine its category, select an appropriate response, or manually send the reply. The client simply receives copies of the outgoing communications for visibility into the system’s activity.
3. How the Automation Works
The application uses a straightforward but effective rules-based approach to identify and respond to customer inquiries.
Step 1: Incoming Email Reception
Customer emails arrive through a mail pipe, which provides the application with access to incoming messages for processing.
Step 2: Subject Analysis
The application examines the subject line of each incoming email and compares it against a collection of predefined keywords and phrases.
Step 3: Automatic Categorization
When a matching phrase is identified, the application assigns the email to its corresponding response category. For example, an email with the subject: “Received wrong item” is automatically assigned to response category 8.
Step 4: Response Selection
The application retrieves the available responses for the identified category from its local SQLite database. Rather than always sending the same message, the application randomly selects one of the available response variations within that category. This provides some variation in the wording of automated communications while maintaining the predefined response content.
Step 5: Default Response Handling
Not every incoming email will necessarily match a predefined subject category. To account for this, the application includes a default response category. When no matching category is identified, the application selects a response from this default category rather than leaving the message without a designated response.
Step 6: Automated Email Delivery
Once a response has been selected, the application sends it to the customer through the client’s dedicated Amazon email account.
Step 7: Client Notification
Every outgoing response automatically includes the client’s dedicated monitoring email address as a CC recipient. This provides the client with visibility into the correspondence being handled by the application without requiring manual involvement.
4. A Fully Unattended Workflow
One of the most important aspects of this project is that the application operates independently. Once deployed and running, it handles the complete email processing and response workflow without requiring employees to initiate individual tasks or make decisions about routine inquiries.
There is no need for someone to:
• Monitor incoming messages.
• Manually review routine inquiries.
• Determine the appropriate response category.
• Select a prepared response.
• Send the response to the customer.
The application handles these steps automatically. The CC functionality provides an additional layer of operational visibility, allowing the client to see the responses being sent while the automation continues to operate independently. This transforms what was previously a recurring manual responsibility into an automated business process.
5. Technical Architecture
The solution was developed using a combination of Go, MySQL, and SQLite. Each technology has a specific role within the application.
Go — The Automation Engine
Go powers the core application and provides the processing logic. It handles incoming email processing, subject analysis, category matching, response selection, and outgoing email delivery. Go was an appropriate choice for this project because of its performance, efficiency, and suitability for long-running background services.
MySQL — Primary Database
MySQL provides the application’s primary database infrastructure.
SQLite — Response Library
SQLite maintains the local repository of response categories and their associated message variations. Separating the response content from the application’s core logic provides a flexible architecture. The response library can be maintained independently without requiring changes to the Go application.
Mail Pipe — Incoming Email Integration
The application retrieves incoming Amazon customer emails through a mail pipe, providing the entry point for the automated workflow.
Dedicated Amazon Email Account — Outgoing Communications
Automated responses are sent through a dedicated Amazon email account.
CC Notifications — Operational Visibility
Every outgoing message is copied to the client’s dedicated monitoring email address, providing visibility into automated customer correspondence.
Amazon Buyer-Seller Messaging Integration
An important aspect of the solution was maintaining the existing Amazon customer communication workflow. The application receives incoming customer inquiries through the client’s email integration and uses the original Amazon-generated encrypted buyer email address when sending responses. Outgoing messages are delivered through a dedicated email account authorized for Amazon Buyer-Seller Messaging.
This allows the automation to operate through Amazon’s established messaging infrastructure rather than communicating directly with customers through unrelated email addresses. Responses are recorded in Seller Central, maintaining visibility into customer correspondence within the Amazon marketplace environment. For additional operational transparency, the application also sends a separate CC copy of every outgoing response to the client’s designated monitoring email address.
6. Designed for Flexibility Without Unnecessary Complexity
An important architectural decision was to keep response categories and message content separate from the application’s processing logic. All predefined response categories and their associated messages are maintained in SQLite. This means that the response library can be modified independently of the Go application. Categories and messages can be added, edited, or removed directly in the database without changing the application code.
A dedicated administrative interface could also be introduced in the future if the client’s requirements warranted it. However, the original solution has proven sufficiently effective that the client has not needed to change its response categories or messages for several years. This is a good example of an important software development principle: build the flexibility a business needs without introducing unnecessary complexity.
7. The Business Impact
The implementation transformed a repetitive customer service responsibility into a fully automated process.
Reduced Administrative Workload
The application handles approximately 150-300 customer emails every day, eliminating the need for employees to manually process routine inquiries.
Faster Response Handling
Subject analysis, categorization, response selection, and email delivery are performed automatically, removing the manual steps that previously delayed the process.
Consistent Customer Communications
Responses are selected from predefined message libraries, helping maintain consistent communication for similar types of inquiries.
Reduced Risk of Delayed Responses
By automating routine email handling, the solution helps the client manage incoming customer correspondence more consistently and reduce the risk associated with delayed or unanswered messages.
Complete Operational Visibility
CC notifications allow the client to see the responses being sent without having to participate in the processing workflow.
Long-Term Maintainability
The application has continued serving the client’s requirements for years without needing changes to its response library or core processing logic.
8. Built to Work. Designed to Last.
One of the most satisfying aspects of this project is its longevity. The application was developed to solve a specific business problem, and it continues to perform that function years after its original implementation. The client receives the benefit of an automated process without having to continually manage the software, adjust response categories, or introduce new systems.
The solution demonstrates that effective automation doesn’t necessarily need to be complicated. By understanding the client’s workflow and developing software specifically for that process, we eliminated a recurring administrative task and created a solution that continues to deliver value.
9. Conclusion
This project is a practical example of how custom software development can make a meaningful difference in everyday business operations. What was once a repetitive manual process requiring employees to handle hundreds of customer emails every month has become a fully automated workflow. By combining Go, MySQL, SQLite, and custom email processing logic, we developed a solution tailored to the client’s operational requirements. The result is less manual work, more consistent customer communications, greater operational efficiency, and a system that continues to work independently years after implementation.
Our philosophy is simple: identify the processes that consume valuable business resources, then build practical, reliable automation solutions that give that time back.
Automation That Makes Business Better
Automation isn’t about replacing meaningful customer service. It’s about eliminating repetitive work so businesses can focus their time, energy, and resources on delivering better customer experiences.
Have a repetitive business process that consumes too much time?
Every business has opportunities to eliminate manual work and improve operational efficiency. We develop custom workflow automation solutions designed around the way your business actually operates. Let’s discuss how automation could help your business work smarter.