Agent / agentic AI

What is agentic AI?

Unlike a simple AI chatbox, Agentic AI are systems designed to pursue goals independently, adapting their actions based on the context and outcomes of previous actions. They can plan multiple steps, use external digital tools (like querying databases, executing software, or calling APIs) independently to achieve pre-determined goals.

Agentic AI examples

• Claude Code – the tool that codes in console or in Visual Studio Code is the classic agent. Unlike basic chat models that just answer questions, Claude Code can plan, write the actual code and self-correct
• OpenClaw – self-hosted AI agent that automates real tasks. It can autonomously manage your email (read, triage, draft, reply via Gmail), handle your calendar and scheduling, work with Google Drive, Docs and Sheets, run Python , process spreadsheets and generate charts. Through integrations it connects to over 1,000 external services — CRMs like Salesforce, Notion, Linear, Slack, GitHub. It can also generate images for social media or marketing, monitor news feeds and deliver briefings, trigger n8n/Zapier automations, and perform web tasks through its built-in browser.