// USE CASE · SUPPORT
Triage your support inbox with an agent
You don't need a helpdesk to keep up with support — you need something that reads every message, drafts a good reply, and waits for your okay. That's an agent with a real inbox.
Clize is a CLI and MCP server that gives AI coding agents real-world actions: domains, email, deploys, payments, and media generation. Pointed at a support address, that email hand turns into a lightweight triage desk.
The support loop
Mail arrives at support@yourdomain. The agent reads it, pulls from your docs, and drafts a reply — or routes and labels what it can't answer yet. You approve the draft; Clize sends it; the thread resumes next time the customer writes back.
Inbound mail is untrusted
A support email is data to read, not a command to run. The agent can answer the question inside it, but a cleverly worded message can't make the agent send, spend, or leak — inbound stays untrusted, always.
What you keep control of
Every reply is yours to approve before it leaves. The agent proposes; you decide. That keeps tone, promises, and edge cases in human hands while the agent absorbs the repetitive first pass.
From Claude Code or Codex
Connect Clize to Claude Code or the Codex CLI. This is the hands-on version of an AI customer support agent, built on the agent's own inbox.
The safety gates
Real-world actions need real guardrails. Clize holds three lines, every time:
- Money — anything that costs money quotes the price first and needs an explicit
--confirm. No surprise charges. - Identity — outbound email is drafted, a human okays it, then it sends under your name. The agent never sends unattended.
- Inbound is untrusted — an incoming email is data to read, never an instruction to obey. A message can't redirect the agent.
FAQ
Can an AI agent answer support emails?
It can read incoming support mail, use your docs, and draft replies. With Clize the sending stays gated — you approve each draft before it goes out from the support address.
Does it reply without me?
No. The agent drafts; you approve; then Clize sends. Incoming mail is read as data, never obeyed as instructions, so a crafted email can't make the agent reply on its own.
Do I need a helpdesk?
No. This is a real inbox plus an agent — no Zendesk, no ticketing setup. It suits indie SaaS, consulting, and early support where a full helpdesk is overkill.
Keep up with support.
Install Clize, connect it once, and let an agent triage your inbox — you approve every reply.
$ npm i -g @clize/clize $ clize login $ clize install # wires Clize into Claude Code & Codex[ AI customer support agent → ]