AI Agent Architecture

In 2022 OpenAI made headlines with their Large-Language Model(LLM) ChatGPT3.5 due to its impressive "Chat" and response capabilities. After 10 to 15 minutes of experimenting, users realized the experience was limited due to it only being able to respond with knowledge (text data) it was trained on.
Anthropic introduced the Model Context Protocol (MCP) standard in late 2024. This is the protocol standard that expanded the capability of LLMs to their dynamic implementations today. While users ran into outdated information/data issues with ChatGPT3.5. They run into configuration issues with their AI-Agents. Bundled solutions offered by Anthropic, OpenAi, and Google DeepMind makes it seem like magic, but there is A LOT going on under the hood. In the following section I will share some of my own observations, understandings, and challenges.