What is EnClaw?

Imagine a colleague who never sleeps, never loses a file, and never minds being asked twice. You brief them in plain words — "chase the overdue invoices from last quarter and draft the reminder emails" — and they go and do the legwork: checking the accounts system, finding the right contacts, writing the drafts. Then, before anything actually leaves the building, they stop at your desk and say: "Here's what I'm about to send. Happy?"

That's EnClaw. It gives you a team of AI assistants — we call them agents — that you talk to the way you'd message anyone else at work: in the web app, or straight from Slack, Microsoft Teams or Telegram. You describe what you want done. The agent works out the steps, uses your organisation's tools to do them, shows you its progress as it goes, and pauses for your sign-off before anything consequential happens.

You don't need to learn commands, write code or memorise magic phrasing. If you can write a clear message to a colleague, you already have the only skill EnClaw asks of you.

Why not just a chatbot?

You've probably used a general-purpose AI chatbot. They're impressive conversationalists, but at work they run into three walls almost immediately.

First, a chatbot can't do anything. It can tell you how one might chase an overdue invoice; it can't open your accounts system, find the invoice, and draft the email. EnClaw agents are wired into the tools your organisation actually uses — email and calendars, JIRA, files, CRM systems and more — so "draft the reminder" means a real draft, addressed to the real customer, ready to go.

Second, a chatbot doesn't know your organisation. It has never heard of your clients, your financial year, or the fact that Priya owns the Acme account now. EnClaw agents work inside your organisation's world, remember what matters, and get better the longer you work with them.

Third — and this is the one that matters most at work — a chatbot has no guardrails. If a public tool could send email, what would stop it sending the wrong thing to the wrong person? EnClaw is built around governance from the ground up: reading information is generally free and instant, but actions that touch the outside world — sending an email, creating a ticket, posting a message — pause for a human yes before they happen. Every action is recorded, so there's always an answer to "what did the agent do, and who approved it?" That combination is what lets your organisation say yes to AI doing real work, not just talking about it.

The shape of a working day with EnClaw

In practice it feels unremarkable, in the best way. You ask a question in Slack and the right specialist answers. You hand over a fiddly task — "pull the open deals closing this month into a summary, owner and value, soonest first" — and watch a single status message update itself as the agent works. Something needs your approval; you glance at the draft, reply "approve, but cc Priya", and it goes. You close your laptop mid-task and the work finishes without you, waiting in the app the next morning.

You stay the decision-maker throughout. The agent does the digging, the drafting and the assembling; you do the judging. And if anything ever heads the wrong way, typing stop halts it.

How this Overview works

The chapters in this section are meant to be read in order, like a short book. They don't teach you which buttons to press — the rest of the guide does that, page by page — they give you the ideas underneath, so the buttons make sense when you meet them.

The natural place to start is with the workers themselves: what an agent actually is, and how EnClaw decides which one should answer you. That's the next chapter.