Before I even start writing this post I would like to apologize because it is about a matter that I really know almost nothing about.
I feel like I have been left myself very far behind, to what comes to AI Agents.
I have been using AI as my assistant on a daily basis, but I have done absolutely nothing about learning more regarding AI Agents.
In this post, I will start from point zero.
Please note that parts of this post have been assisted with information coming from AI and also as included in the sources at the end
Everything starts with a definition.
What is an AI 🤖 Agent?
An AI agent is a program that can observe → decide → act to achieve a goal.
To put it simply: An AI agent is an AI that can do tasks for you, but in an automated way.
image screenshot by flowiseai.com
When we use a chatbot, it is programmed to simply answer our questions. The AI agent though, can receive input in the form of data, message, tasks and then process it (decide) using AI in order to take the action we have instructed (it can be searching the web, send messages, write posts, analyze data etc)
This process is repeated until the goal is accomplished.
Some examples could be: A social media AI agent
Reads crypto news
Writes a post
Publishes it automatically
Or A customer support AI agent
Reads a customer question
Finds the answer
Replies automatically
I have asked what is the best way to start and I will be testing two free ways.
The easiest - no code way for now is Zapier.com/ai and I will also test out the free option of https://flowiseai.com/
... To be continued....
If you have created an AI Agent, or if you have any link/video to share with me as I am starting now, I will be grateful!
Thank you!
More info on AI Agents:
https://cloud.google.com/discover/what-are-ai-agents
https://howaiworks.ai/glossary/ai-agent
I tried openclaw a few weeks ago. It's pretty impressive at first and I see why there's so much hype around it... in the end it just a more versatile version of claude code.
I think it's importand to understand that "agentic ai" is still just an LLM predicting output just wrapped into an interface that provides context and tool calling in an "agentic loop".
It is pretty powerful, but it's still very prone to the usual LLM shenanigans, which I would describe as overconfidence. I found myself micromanaging my openclaw assistant a lot... but maybe that's also a trust issue on my side :P
and all the compute such an assistant needs is not to be underestimated... after a few hours of playing with openclaw I had burned several million tokens with anthropic!
Thanks for sharing!
i don't use ai during the day, even worse an ai agent 😂