Docs · Product

AI agent

A helper that answers FAQs from your knowledge base while you sleep. Hands off to you when it's out of its depth. Two checkboxes and a system prompt — that's the whole config.

How it works

When a visitor sends a message, the AI receives the conversation so far, the relevant chunks from your knowledge base, and a short system prompt that establishes tone. It posts a reply marked with an AI Assistant badge so the visitor can tell at a glance they're not talking to a person yet.

If a question is outside the AI's confidence, the conversation is flagged as waiting for a human and the on-shift operators get a notification.

Turning it on

The AI is off by default. To enable it:

1Open Settings → AI Agent.
2Flip the Let the AI answer customers toggle to on, then click Save.
3The next inbound visitor message is handled by the AI.

Tip

Nothing else is required — no API keys, no model picking, no provider setup. Leadiosa manages the LLM under the hood; you just decide when it speaks.

When the AI helps

Two checkboxes on the same settings page decide when the AI steps in:

Answer customers when nobody from my team is online — covers after-hours, weekends, lunch breaks. The AI posts a public reply so visitors don't bounce.
Suggest replies for me when I'm online — a draft appears above your reply box. One click to send, or edit first. Nothing is posted to the visitor until you choose.

Pick one, both, or neither. Both off is functionally the same as the master toggle being off.

How the AI talks about your business

One textarea on the settings page where you tell the AI what it is. A few sentences is enough — your business name, the tone you prefer, things to never promise, where to send people who want to book.

Example for a yoga studiotext

You're answering on behalf of Lotus Yoga Studio in Brooklyn.
Friendly, concise, never push hard. Direct booking questions
to lotusyoga.example/book. If someone asks about refunds,
say you'll forward it to the studio owner.

The AI also reads your knowledge base on its own (see below) — you don't need to repeat FAQ answers here.

Per-conversation toggle

Want the AI off for a particular conversation but on everywhere else? Open the conversation, click AI · on in the header — it flips to AI · off for that thread only. The workspace default is unaffected.

Knowledge-base grounding

The AI retrieves the most relevant chunks from your knowledge base on every turn and includes them in the prompt. Replies cite the source articles by title.

text

Visitor: Where do I export invoices?
AI:      Invoices are exportable from Settings → Billing →
         Invoices. Click "Export CSV" to download the
         current quarter.
         (Source: "Invoices and exports")

For the AI to actually help, give it something to read — see Knowledge base.

Monthly AI use

Every AI reply consumes tokens. The settings page shows how many you've used against your plan's monthly allowance — read-only, included with the plan, no need to configure anything.

When the allowance runs out, the AI pauses until the next calendar month. Visitor messages still arrive in your inbox and your team answers as normal. To raise the limit, upgrade the plan.

Under the hood (for the curious)

The model and provider are managed by Leadiosa, not exposed per workspace. We route through OpenRouter to OpenAI's GPT-class models by default — the provider tier is contractually excluded from training on customer content. If we change models, replies just keep working; nothing in your workspace settings has to update.

Embedding for knowledge-base retrieval is similarly platform-managed. See Knowledge base for the data-flow details.

PreviousOperator inbox Next Knowledge base