> ## Documentation Index > Fetch the complete documentation index at: https://docs.orbitagents.xyz/llms.txt > Use this file to discover all available pages before exploring further. # Chat (create with Orbie) > Describe an outcome in plain English and Orbie builds the agents, picks the tools, and files the work — the fastest way to hire an AI team. Chat is the front door to Orbit — and the primary way to create almost anything. You describe an **outcome**, and **Orbie** (your Orbit assistant) figures out the rest: it designs the agents, picks the right tools, and files the work where it belongs. It's the closest thing to hiring — without hiring. ## Describe the outcome, not the steps You don't configure agents field by field. You say what you want: > "Find companies matching my ICP every morning, enrich the decision-makers, and drop them into a tracker." > "Watch my competitors' sites and pricing weekly and write me a short brief on what changed." > "Read what's in my Vault each week and draft posts in my voice, ready for me to review." Orbie thinks about the rest — which agents to create, which tools they need, how they should run — and sets it up. You review and approve. ## It replaces hiring Instead of hiring for a role, you describe the role and Orbie builds the agent that fills it: * **GTM engineer** — lead sourcing, enrichment, signal watching, wired into your tools. * **GTM / growth marketer** — research → content → outreach, on a schedule. * **Social media marketer** — drafts and posts in your voice across channels. * **Automations** — several agents chained into one workflow (see below). One person, describing outcomes, ends up running the output of a whole team. ## It can build automations, not just single agents The revolutionary part: Orbie doesn't only make one agent — it can wire **several agents into an [automation](/concepts/automations)** where each works on the previous one's output, in sequence. A research agent feeds a writer, which feeds a publisher. You describe the end-to-end outcome; Orbie builds the chain. This is the cross-coordination that's been missing from AI and agent development: agents that hand off to each other, all sharing one memory, all governed by your decisions. ## It also answers anything about your setup Chat isn't only for creating. Ask it about your own data and workforce: * "What did my agents produce this week?" * "Which leads did we contact and what happened?" * "What's in my Vault about pricing?" Because Orbie sits on top of your Vault, decisions, and every agent's output, it can answer from *your* company's real data — not generic knowledge. ## What you'll see * A box that says **"Describe what you want to create."** * Chips to scope it: **Project**, **Department**, **Add context**, a **model** picker, and **History**. * Suggested starting points like *Build agents to grow my business*, *Help me set Orbit up*, *Watch my competitors*, *Turn research into content*, *Find me new leads*. ## Chat vs. the other create paths | Path | Best for | | ---------------------------------------- | ---------------------------------------------------------------------------- | | **Chat (Orbie)** | Describe an outcome; Orbie builds agents or a whole automation. The default. | | **[Launch an Orbie](/use/launch-orbie)** | Grab a ready-made agent in one click. | | **[Agent Builder](/use/agent-builder)** | Hands-on build of one agent from a goal. | All three land the same place: real teammates in your AI company that share your memory and report to you. Chain agents into an automation. The bigger picture for solopreneurs.