Even in 2025, many companies rely on direct customer support – and in most cases, this still happens via email. It’s no surprise: emails are fast, flexible, and indispensable in any business. However, they are also a massive time sink. Especially in sales, the manual effort quickly adds up to several hours a day – time that could be better spent on customer relationships or strategic topics. This is where there is enormous potential for optimization that can be easily tapped into using modern language models (LLMs). Automating email-based processes is one of the areas where LLMs can demonstrate their strengths particularly well – as we have already shown in a previous edition. Comsysto Reply has been implementing complex e-commerce solutions for clients across all industries for over 15 years. In parallel, we analyze trends and best practices in our e-commerce lab. We share our insights here – every other Tuesday!
System integration as the key to success
The biggest challenge often lies not in the AI itself, but in the integration: for automated process support, a wide variety of information from CRM, ERP, and ticket systems must be brought together. Only when this connection functions smoothly and the established business process is optimally supported can the full potential of AI-powered email automation be realized.
Between standard software and customized solutions
To efficiently master this balancing act, we develop so-called reference applicationsin the Comsysto e-Com Lab. These contain the core logic of a specific business case – such as email automation in sales – and serve as the foundation for customer-specific projects. Building on this, only the individual extensions and integrations are implemented – such as adapters to existing systems, workflows, and user interfaces. This creates a solution that is ready for use quickly while still fitting the respective company perfectly.
Use case: Email automation in sales
The current reference application focuses on exactly this use case. The goal is for the AI to take over as much routine work as possible, while the final decision remains with the human. Business users can review the results and make adjustments as needed – this minimizes manual steps while simultaneously building trust in the automation step by step.
This is what the future of email processing looks like
A screenshot from our prototype application shows how incoming emails are automatically classified. Each inquiry is broken down into its various requests – for example, quote requests, delivery status, or complaints – and assigned to a specialized AI agent.
As a result, the reference implementation can multiple requests within a single email identify and, for each request, provide optimized support and a specific user interface . This allows even complex issues to be handled in a structured way, without having to fall back on purely manual processing.
The result: Humans retain control, while AI handles routine tasks and context preparation – together creating significantly more efficient and higher-quality customer communication.
That’s it for today – and here’s what’s next
In our next issue, we will return to a key architectural aspect of modern enterprise software: Composable Commerce – and how modular systems can further increase innovation and efficiency in sales.
See you in 2 weeks!
Christian & the Comsysto Reply e-Commerce Lab