Build AI assistants, for every team, without code.
With RAGSuite you build your own AI assistants for HR, sales, IT or customer service. Each assistant gets its own sources, instructions and language model, and names the source on every answer. Done in minutes, in the browser.
What is an AI assistant, and what is an AI agent?
An AI assistant is a chat with a fixed task: it answers questions on one topic from exactly the sources you give it, and follows your instructions. An AI agent goes one step further and carries out tasks itself, for example creating a ticket.
RAGSuite assistants answer questions with sources. If an assistant should trigger something, you connect it to your systems via n8n, webhooks or MCP, so you stay in control of what happens automatically.
Five steps to your own AI assistant.
Define the task
Who is the assistant for, which questions should it answer? A clear purpose makes good answers.
Choose sources
Files, website sections or connectors such as SharePoint and Confluence, only what belongs to the task.
Write instructions
Tone, length, language and limits: for example “refer unclear cases to the HR team”.
Choose a model
A small model for simple questions, a large one for hard ones: set per assistant.
Test and share
Ask test questions, check the answers, then share with a team or embed as a widget.
Six AI assistants teams build first.
HR assistant
Leave, travel expenses, parental leave, onboarding, from the handbook and works agreements.
Contract assistant
Find deadlines, termination clauses and price rules in framework contracts.
IT help desk
VPN, passwords, software requests, from the IT wiki, around the clock.
Website chatbot
Customer questions on products, delivery and service, from your website.
Sales assistant
Product data, price lists and objections, for quotes and calls.
Onboarding assistant
New colleagues find processes, contacts and tools.
AI assistant with RAGSuite, a general chatbot or your own build?
| Criterion | General AI chatbot | Own build | AI assistant with RAGSuite |
|---|---|---|---|
| Knowledge | Training data, files per user | What your team connects | Your sources, per assistant |
| Source on every answer | Usually not | If built | ✓ |
| Effort | Low | Weeks to months | Minutes, no code |
| Roles and access | Per account | If built | Teams and roles |
| Where data lives | With the provider | With you | RAGSuite Engine or your server |
Who may use which assistant?
Each assistant belongs to a project with its own sources. In the RAGSuite Engine and in Enterprise you share assistants with teams and roles, Google sign-in included. The audit log shows who asked what; deletion periods and switchable storage protect your staff.
Access rights from SharePoint or Confluence are not carried over automatically. So only add sources to an assistant that all its users may see.
Build your first assistant today.
Create an account in RAGSuite Engine, upload a handbook, write an instruction, ask. 7 days free, no credit card.
Questions about building AI assistants.
How do I build an AI assistant?
You create a new assistant in the browser, give it a task, choose its sources, write short instructions and pick the language model. Then you ask test questions and check the answers and their sources. If all is well, you share it with a team or embed it as a widget. No coding is needed.
What is the difference between an AI assistant and an AI agent?
An AI assistant answers questions from set sources and follows your instructions. An AI agent also carries out tasks itself, such as creating an entry. RAGSuite assistants answer with sources; you trigger actions through n8n, webhooks or MCP. That keeps it clear what happens automatically.
How many assistants can we build?
As many as you need. Each assistant belongs to a project with its own sources, instructions and settings. Projects are unlimited in the Community Edition. In the RAGSuite Engine you pay the plan price, not per assistant; all assistants share the credits of your workspace.
Which sources can an assistant use?
Uploaded files such as PDF, DOCX and TXT, crawled websites, and content from SharePoint, Confluence, Google Drive, Notion, Slack and Gmail. You choose only the sources that fit each assistant’s task. Access rights from the source systems are not carried over automatically, so you separate content by project.
Can we use an assistant on our website?
Yes. Every assistant can be embedded as a chat widget with one snippet in your website, intranet or CMS, in your colours, with a greeting and only on the domains you allow. Users see a notice in the widget that they are talking to an AI. The widget works in the RAGSuite Engine and self-hosted.
Where does our assistants’ data live?
In RAGSuite Engine, sources, search index and history live in a data centre in Germany. Self-hosted, everything stays on your servers. If you use an external language model, only the question and the matching passages go to that provider. With a local model, self-hosted, nothing leaves your network.
Build your first AI assistant.
7 days free in the RAGSuite Engine, or we set up the first assistant together in a demo.