{"id":107943,"date":"2026-10-22T11:04:10","date_gmt":"2026-10-22T11:04:10","guid":{"rendered":"https:\/\/mailrelay.com\/glossary\/ai-agent\/"},"modified":"2026-10-22T11:04:13","modified_gmt":"2026-10-22T11:04:13","slug":"ai-agent","status":"publish","type":"glossary","link":"https:\/\/mailrelay.com\/en\/glossary\/ai-agent\/","title":{"rendered":"AI agent"},"content":{"rendered":"\n

An AI agent (or artificial intelligence agent) is a system that receives a goal, decides which steps to take to achieve it and uses external tools to carry them out. It doesn’t just answer: it looks up data, performs tasks and checks the result.<\/p>\n\n

For example, a chatbot explains how to create a segment. An agent goes into your email marketing tool, creates the segment and tells you how many contacts it has.<\/p>\n\n

How does an AI agent work?<\/h2>\n\n

Behind almost every agent there is a language model, like the ones used by ChatGPT or Claude. What turns it into an agent is everything around it.<\/p>\n\n

It usually works in a five-step cycle:<\/p>\n\n

    \n\n
  1. Goal.<\/strong> You tell it what you want to achieve, in your own words.<\/li>\n\n\n\n
  2. Plan.<\/strong> The agent breaks the task down into smaller steps.<\/li>\n\n\n\n
  3. Tools.<\/strong> It uses the applications it has access to: an API, a search engine, a spreadsheet or your email marketing account.<\/li>\n\n\n\n
  4. Memory.<\/strong> It keeps track of what it has done and what it has found to decide on the next step.<\/li>\n\n\n\n
  5. Review.<\/strong> It checks the result and, if something goes wrong, it adjusts the plan or asks you for confirmation.<\/li>\n\n<\/ol>\n\n

    Tools are the key. A model with no connections can only write. With connections, it can act. Today, the most widespread way to provide those connections is MCP (Model Context Protocol)<\/a>, an open standard used by the leading AI assistants.<\/p>\n\n

    \"How<\/figure>\n\n
    \n

    Connect your email marketing to AI<\/strong><\/p>\n\n\n\n

    Create your free account and request access to the beta of the Mailrelay MCP server.<\/p>\n\n\n\n

    \n
    Create a free account<\/a><\/div>\n<\/div>\n<\/div><\/div>\n\n

    What is the difference between a chatbot, an assistant and an AI agent?<\/h2>\n\n

    All three use artificial intelligence, but they do different things.<\/p>\n\n

    <\/th>Chatbot<\/th>AI assistant<\/th>AI agent<\/th><\/tr><\/thead>
    What it does<\/td>Answers questions from a script or a database<\/td>Converses, writes and summarizes<\/td>Achieves multi-step goals<\/td><\/tr>
    Uses tools<\/td>No, or very few<\/td>Sometimes, if you give it them<\/td>Yes, it’s what it’s built for<\/td><\/tr>
    Decides the steps<\/td>No<\/td>Only within the conversation<\/td>Yes, it plans and corrects<\/td><\/tr>
    Example<\/td>FAQ bot on a website<\/td>Asking it to write a subject line<\/td>Asking it to prepare and schedule a campaign<\/td><\/tr><\/tbody><\/table><\/figure>\n\n

    The line is blurring. Many assistants already work as agents when you connect them to your applications. Even so, the distinction helps you know what to expect from each one.<\/p>\n\n

    What types of AI agents are there?<\/h2>\n\n

    Single-task agents.<\/strong> They do one specific thing, such as sorting incoming emails or answering first-line support tickets.<\/p>\n\n

    Workflow agents.<\/strong> They chain several tasks within a process: finding a piece of data, preparing a document and sending it for review.<\/p>\n\n

    Agents connected to your tools.<\/strong> They work inside your applications (CRM, online store, email marketing) with the permissions you give them.<\/p>\n\n

    Multi-agent systems.<\/strong> Several specialized agents share the work. One researches, another writes and another reviews.<\/p>\n\n

    Examples of AI agents in marketing<\/h2>\n\n