Sebati Guide

Welcome

The technical guide to using Sebati, for the people who work with it and the people who build with it.

What is Sebati?

Sebati brings the most intelligent AI into the way your organization works. Agents are shaped around your knowledge, systems, workflows, and approvals, so AI becomes part of how work actually happens. Your deployment runs on your own infrastructure and carries your organization's name, rules, and data.

Sebati is also built with you, not just sold to you. Your Sebati team (forward-deployed engineers, FDEs) helps discover the right use cases, build the first agents, and improve them after launch. That is why this guide covers both using Sebati and building with it, and why neither requires touching source code: everything happens through the product.

Sebati's Chat surface

Examples throughout this guide come from a live demo deployment, using Ava, an executive copilot agent grounded in the demo company's analytics warehouse. Your own deployment will have its own name, branding, and agents; the mechanics are the same.

Five words carry most of this guide:

TermMeaning
AgentPurpose-built AI for a kind of work, with its own knowledge, tools, and boundaries
CaseA structured work surface where agents process rows of repeatable work
ChannelWhere work reaches an agent: chat, WhatsApp, forms, webhooks, schedules
KnowledgeYour trusted documents, ingested so agents answer from them with citations
SkillA reusable procedure that teaches agents how your team works

Start here

Find your path

I want toRead
Chat with agents, attach files, share resultsEveryday use
Create documents, slides, and spreadsheetsCreating content
Let an agent run on a scheduleScheduled work
Configure an agent's prompt, models, and featuresAgents
Give agents trusted company knowledgeKnowledge
Connect Slack, WhatsApp, email, forms, or webhooksChannels
Turn a repeatable workflow into structured AI workCases
Build agents from Claude Code or CodexCopilot MCP
Call Sebati from my own systemsAPI essentials
Manage members, roles, and keysWorkspace administration

Working in a team? A Workspace is where a team's agents, knowledge, Cases, and channels live; your organization can have several, and everything in this guide happens inside one.

On this page