Interface tour for agencies

Chatfuel keeps the interface small on purpose: a handful of day-to-day screens, one page where all of the AI's behavior is configured, and an AI Co-Worker you can hand the same work to in chat. Nothing here needs a dedicated bot builder to operate — but a new team still burns its first hour looking for buttons, so here's the map.

Fuely AI → Automations: the AI's control panel

Everything about where and how the AI replies lives on one page: Fuely AI → Automations — here's the full guide to how it works. Three layers to know:

• Main bot, at the top. Its settings are the defaults for everything, and its toggle is the master switch — Main bot off means no AI replies anywhere.
• A base automation for every traffic source the client's channels expose: Instagram DMs, post & reel comments, ad comments, story replies, Facebook Messenger and post comments, WhatsApp DMs and ads, TikTok, the website widget. Enabling a source's automation is what turns the AI on there, and each one carries its own settings — when the AI replies, its instructions, handoff rules, follow-ups — inherited from Main bot until you override them.
• Custom automations on comment, ad, story, and link sources: named rules targeted at keywords or at specific posts and ads. For an agency this is the per-campaign lever — one client's launch ad gets its own funnel and its own replies, the evergreen campaign gets another, and the two never collide.

That layered structure is why one Chatfuel setup can serve very different client campaigns at once. Its sibling page, the Knowledge Base, holds what the AI knows — the client's business facts; Automations decide where and how it behaves.

The AI Co-Worker

The Co-Worker is a chat that can read and change everything the dashboard can — the automations above, the Knowledge Base, reports. Instead of clicking through settings, you describe the outcome in plain language:

• "Set up auto-replies to Instagram comments on our latest ad"
• "Collect the customer's budget and phone number before booking"
• "Create a follow-up for leads who go quiet after 24 hours"

It configures the automation, asks for anything it needs from you, and confirms what it set. The two paths are equals: what the Co-Worker writes, you can open and inspect in Automations; what you change in Automations, the Co-Worker sees. Mixing them is the normal workflow — rough setup in chat, fine-tuning on the page, or the other way around. Use whichever is faster for the change in front of you.

One phrasing habit makes the chat path work noticeably better: describe the outcome, not the mechanics. "I want the AI to qualify leads by asking about budget" gets you further than "add a rule that sends a question." The Co-Worker thinks in results.

For an agency this changes the economics of setup: a junior account manager can configure things that used to require a dedicated bot builder — ask for them in chat, then check the result on the Automations page. What the AI actually runs from — its modules and where each one lives — is covered in Bot modules and blocks, and the knowledge behind it is built by uploading the client's business info.

The day-to-day tabs

Around those two sit the screens your team lives in once a client is running:

And one control that matters more to you than to any single-business user: your avatar → My Companies, the switcher between client companies. Every tab above shows one client at a time — whichever company you're currently in. The full logic of that structure is in Agency → clients model.

Where each of your people spends their day

Give each person the role that matches — the interface they see trims itself to what their role can do, which is exactly what you want when the client's own staff is logging in.