🎧 Prefer to listen?

Last week I asked my AI assistant to check my calendar, find a gap, and schedule a meeting. It did it. No copy-pasting. No switching tabs. Just… did it.

That’s tool calling. And if you’re using AI without it, you’re using 10% of what these models can actually do.

What is tool calling?

When you talk to ChatGPT or Claude, it can only do one thing: generate text. That’s it. It can’t check your email. It can’t search the web. It can’t post to social media. It’s a very smart autocomplete.

Tool calling changes that.

Here’s how it works:

  1. You ask the AI something that requires an action (like “check my email”)
  2. The AI recognizes it needs a tool
  3. The AI sends a request to a connected tool (like your email provider)
  4. The tool does the work and sends the result back
  5. The AI reads the result and gives you the answer

The AI doesn’t do the work. It decides WHICH tool to use, sends the right parameters, and interprets the result. The tool does the actual work.

Think of it like this: The AI is the brain. The tools are the hands. The brain decides what to do. The hands do it.

Why this matters for you

Without tool calling, AI is a chatbot. With tool calling, AI is an assistant.

What you can do with tool calling:

  • “Search the web for the latest news on X” → AI searches, summarizes, gives you the answer
  • “Check my email and summarize the important ones” → AI reads your inbox, filters, summarizes
  • “Post this to Twitter and Instagram” → AI handles both, different formats, right now
  • “Book a flight to New York next Tuesday” → AI searches flights, finds the best option, books it
  • “Analyze this spreadsheet and create a chart” → AI reads the data, creates the visualization

The difference: Without tool calling, you copy-paste between apps. With tool calling, you just ask.

MCP — the standard that connects everything

There’s a new protocol called MCP (Model Context Protocol) that’s becoming the standard way to connect AI to tools.

Before MCP, every AI tool had its own way of connecting. ChatGPT had plugins. Claude had integrations. Every setup was different.

MCP creates one standard. One way to connect any AI to any tool. It’s like USB-C for AI — one connector, everything works.

What MCP does:

  • Exposes tools as “functions” the AI can call
  • Handles the connection between the AI and the tool
  • Standardizes how tools describe themselves so the AI knows what each one does
  • Works across different AI models (ChatGPT, Claude, Gemini, local models)

Why this matters: You don’t need to learn a new system for each AI. Connect your tools once, use them with any AI.

The three layers (simplified)

There are three things happening when AI calls a tool:

Layer 1: Function Calling (inside the AI) The AI model itself has the ability to “call functions.” This is built into models like GPT-4, Claude, and Gemini. When you ask something that needs a tool, the model generates a structured request (not text — a function call with parameters).

Layer 2: MCP (the connector) MCP is the protocol that connects the AI to the actual tools. It takes the function call from the AI, routes it to the right tool, and sends the result back. Think of it as the telephone line between the brain and the hands.

Layer 3: A2A (agent-to-agent) When you have multiple AI agents that need to work together, A2A (Agent-to-Agent) protocol lets them communicate. One agent handles research, another handles writing, a third handles publishing. They coordinate through A2A.

You don’t need to understand all three layers. Just know: Function calling = AI decides what to do. MCP = AI connects to the tool. A2A = multiple AIs work together.

Real examples you can use today

Example 1: Research + write + publish

I asked my AI: “Find the latest research on magnesium and sleep, write a short summary, and save it to my notes.”

The AI:

  1. Called a web search tool → found 5 recent studies
  2. Read and summarized the findings
  3. Called a file tool → saved the summary to my notes

Total time: 30 seconds. Without tools, this would take 20 minutes.

Example 2: Social media management

I told my AI: “Schedule a post about today’s blog article on Twitter at 9am and Instagram at 12pm.”

The AI:

  1. Called the Twitter tool → created a draft with the right format
  2. Called the Instagram tool → adapted the content for Instagram’s format
  3. Scheduled both for the right times

Total time: 10 seconds. Without tools, I’d log into two platforms, format two posts, set two schedules.

Example 3: Email triage

I asked: “Check my email and tell me what needs my attention today.”

The AI:

  1. Called the email tool → read 47 unread emails
  2. Filtered out newsletters, promotions, and automated messages
  3. Summarized the 6 that needed actual responses

Total time: 15 seconds. Without tools, I’d spend 10 minutes scrolling.

How to start using tool calling

If you use ChatGPT:

  • Go to Settings → Connected Apps
  • Add your tools (Google Calendar, email, etc.)
  • Now ChatGPT can use them in conversations

If you use Claude:

  • MCP servers can be connected in Claude’s desktop app
  • Add tools one by one
  • Each tool becomes available in your conversations

If you want full control:

  • Use an open-source agent framework (like n8n, LangGraph, or OpenClaw)
  • Connect your own tools
  • Run everything locally

The last option is the most powerful but requires more setup. The first two are the easiest to start with.

The bottom line

AI without tool calling is like a calculator that can only show you the formula but never compute the answer. It’s smart but useless.

Tool calling turns AI from a chatbot into an assistant. The AI thinks. The tools do. And you just ask.

If you’re not using tools with your AI yet, start today. Connect one tool. See the difference. Then connect more.

The era of “AI as a chatbot” is over. The era of “AI as an assistant” just started.

Start with one tool. I recommend connecting a web search tool first — it’s the most immediately useful. Then add email or calendar. Build from there.

Coming soon on No Code Required:

  • Postiz: the free open-source replacement for Buffer and Hootsuite
  • AiToEarn: the free scheduling tool that actually pays you to post
  • Google’s $100/month AI plan — is it worth it?