Selzy MCP Server Lets AI Assistants Run Email Campaigns

What happened

Selzy has launched a Model Context Protocol (MCP) server that connects AI assistants directly to its email marketing platform. Small businesses can use it to prepare, launch and analyse email campaigns through a conversation with an AI assistant, according to CustomerThink's report.

The company's internal testing put a concrete figure on the change. Preparing a campaign took around 20 minutes before the integration and about five minutes with it. The announcement positions the tool as a way for a small team, or a single person, to handle work that would usually be spread across several roles.

Why it matters

This is less about a new email feature and more about a new interface. When an AI assistant can act on a live marketing platform, the operator's job shifts from navigating menus and forms to stating an intent and reviewing the result. That is a meaningful change for organisations that have historically needed specialist marketing staff to run campaigns, and it shows how AI adoption is moving from drafting content into executing operational tasks.

For service and digital leaders, the wider signal is that integration standards such as MCP may matter as much as the underlying model. Tools that expose their capabilities cleanly to AI assistants could become easier to adopt inside small and mid-sized businesses, where the tooling and workflow decisions are made quickly and by few people.

By the numbers

• 20 minutes to prepare a campaign before the integration, according to Selzy's internal testing as reported by CustomerThink.
• Five minutes to prepare a campaign with the MCP server in place, a reduction of roughly three-quarters.

The Renascence take

Most coverage will focus on the time saving, but the more important question is what happens to judgement when the conversation becomes the workflow. A five-minute campaign is only an improvement if the brief, segmentation and review steps are still done properly.

The real design challenge is not whether an assistant can build an email in five minutes; it is whether the owner can still see what is being sent to whom and why. Operators should treat the conversational layer as a new interface that needs its own checkpoints, such as explicit confirmation before a send, plain-language summaries of audience and timing, and a record of what the assistant changed. Businesses that map these moments before rollout will get the speed without the silent errors that come with it.