In almost every meeting about artificial intelligence there comes a moment when someone says “agent”, someone else hears “chatbot” and a third person thinks of automation. The conversation goes on, everyone agrees, and each person leaves the room with a different project in mind.
Nobody lacks intelligence. What is missing is a shared vocabulary.
AI gained new words faster than we could learn them.
Token, context window, RAG, fine-tuning, hallucination, MCP, harness, evals. In little more than two years, terms that lived in academic papers and engineering forums started showing up in sales proposals, board presentations and campaign briefs.
Many arrive with no explanation. Others arrive with explanations that confuse more than they clarify. And they almost always arrive without context: nobody says what they are for, where they tend to go wrong or why they should matter to people who do not write code.
The result is predictable. Important decisions end up being made on the basis of words that each person understands differently.
Vocabulary is infrastructure too.
Anyone who has worked on a communication project knows that a large part of the job is aligning meaning. What is positioning? What counts as a conversion? What exactly do we mean by brand?
AI is no different, just newer. Knowing the difference between training a model and giving it context changes a project’s budget. Understanding what a hallucination is changes the kind of review a piece of content needs. Knowing what an agent can do on its own changes the conversation about risk.
You do not need to become a specialist. You need to be able to ask the right questions.
That is why we built a glossary.
The Media Clorians AI Glossary brings together 313 terms, from algorithm to agent, organized into seven topics: fundamentals, neural networks and training, LLMs and language, prompts and context, models and market, agents, and quality and risks.
Each entry has a short definition, an analogy to make the idea stick and, when it makes sense, a warning about what is commonly misunderstood. Terms link to one another, and each one lists its sources.
You can search by the English or the Portuguese name, filter the 54 essential terms for beginners or see only what emerged between 2024 and 2026.
A glossary does not replace the conversation. It makes the conversation better.
AI will keep changing fast, and the glossary will change with it. But the idea behind it stays the same: before deciding what to do with a technology, it pays to agree on what the words mean.
If, while reading an entry, you think “so that is what they meant in that meeting”, the glossary has done its job.
