{
  "id": 19150514,
  "title": "Build Your First AI Agent",
  "url": "https://urgent.news/2026/11/09/build-your-first-ai-agent",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-09T12:27:01.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/paragonai/build-your-first-ai-agent-87b"
  },
  "original_language": "en",
  "account": "OpenAI is ending Agent Builder, the drag-and-drop tool for creating agents within ChatGPT, on November 30, 2026. This episode emphasizes the importance of building agents with your own tools rather than relying on a vendor's platform, as agents built on vendor tools can be abruptly shut off.\n\nTo build your first AI agent, follow these five steps:\n\n1. Define a single job as a clear outcome. For example, \"Monitor the price of this flight and email me if it drops below $300.\"\n2. Choose 1-3 tools, such as email, calendar, or web search. Fewer tools reduce the chances of failure.\n3. Implement basic memory for the current conversation.\n4. Include a trigger with a strict approval step, where the agent drafts an action, and you approve before any real action occurs.\n5. Use a loop that follows the sequence: see, think, act, check, and repeat until the goal is achieved or human intervention is required.\n\nWhen building an agent, keep the number of tools low. If an agent takes five sequential steps, and each step has a 90% reliability, the overall success rate drops to only 59%. This highlights the importance of short jobs and minimal tools.\n\nNo coding is required to build your first agent. Compare visual builders like Zapier Agents, Make, and n8n in terms of ease of use and pricing. Zapier Agents are the easiest for non-technical users, with over 9,000 integrations and 400 free activities per month, costing $33.33 per month for the Pro plan. Make offers 1,000 operations free, with a Core plan at $9 per month for 10,000 credits. However, Make charges per full workflow run, not per step, making it more expensive for complex agents.\n\nCustom GPTs within ChatGPT are a fast way to create drafting and research agents, but they rely on OpenAI's platform, which will soon be shutting down Agent Builder. Other builders like Relevance AI, Gumloop, and Lindy can be used when the big three are not suitable.\n\nThe episode demonstrates building a price-drop watcher agent in six steps. The job is to monitor a flight or product price and email you when it drops below a specified amount. The two tools needed are a price check and an email notification. The agent is tested with five real examples before being deployed. Guardrails are set, such as read-only mode, a spending cap of zero, and a separate test account for any write operations.\n\nThe episode concludes by highlighting that 40% of agent projects fail due to escalating costs, unclear business value, and weak risk controls, according to Gartner's 2025 prediction. Despite this, only 17% of organizations have actually deployed agents, even though 60% expect to do so within two years.",
  "summary": "Paragon AI MAG, Series 2, Episode 2 — \"AI Agents: The AI That Does the Work\" OpenAI is shutting down Agent Builder — the drag-and-drop part of AgentKit for building agents inside ChatGPT — on 30 November 2026. Anyone who built an agent on it has weeks to rebuild somewhere else, or their agent simply stops working. That is the whole lesson of this episode in one sentence: an agent you build on a…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}