Assistants

Build reusable AI assistants with their own persona, knowledge, tools and behaviour – managed centrally in the Library and launched into chat with a single click.

Assistants (also labelled Agents in the interface) are persistent chat personas. Instead of retyping the same instructions for every conversation, you build an assistant once – with a fixed system prompt, a preferred model, its own knowledge and the right tools – and then launch as many chats from it as you like with a single click. All assistants live centrally in the Library and are visible company-wide.

What an assistant is

An assistant bundles several things into one reusable profile:

a persona (system prompt) that steers behaviour in every chat
a preferred model and model parameters (creativity, tokens, reasoning)
its own knowledge (attached documents) used automatically when answering
tools via integrations (e.g. web search, OneDrive, Google Drive)
conversation starters and a greeting message for a smooth start
optional subagents and skills for more complex tasks

ℹ️

Assistants are company-wide. By default every member of your company can see an assistant and use it in chat. Only the creator, workspace admins/owners and colleagues with an explicit Edit share may modify it.

The Library

You reach assistants via Library → Agents. The page has three tabs:

Agents – every assistant in your company
Workflows – multi-step, form-based flows (a separate module)
Templates – ready-made templates to copy

The Library Agents tab: search, filter and launch a chat with one click.

Searching and filtering

In the Agents tab, the toolbar offers:

a search box (matches name and description)
the filter pills All, Mine (only those you created) and Pinned
a label bar: click one or more labels to filter by topic (OR logic – a single match is enough)

Highlighted assistants automatically float to the top of the list.

An assistant card

Each card shows the icon (emoji on the brand colour), the name, a two-line description, up to three labels, and in the footer the model plus a Web badge if web search is on. A shield icon marks verified assistants.

Click the card → immediately starts a new chat with that assistant.
Pencil icon → opens the editor (only visible to the creator and admins).
Trash icon → deletes the assistant (creator and admins only; irreversible).

Creating an assistant

Go to Library → Agents.
Click + New agent in the top right. A blank assistant ("New agent") is created immediately and the editor opens.
Fill in the sections (see below) and save.
Click Publish so your configuration takes effect in real chats.

💡

Start faster from a template: switch to the Templates tab, filter by category and click Copy to my company on a suitable template. You land directly in the editor with pre-filled values that you can adjust freely.

The editor

The editor is split in two: configuration on the left, a live preview on the right (on large screens) where you can try the assistant out instantly. The header bar offers the actions ▶ Test, Clone, Save and Publish, plus an Unpublished changes indicator when relevant.

The editor: configuration on the left, live preview on the right.

Identity

Avatar: upload a custom image via Image (PNG, JPEG, SVG or WebP, max. 2 MB). An uploaded image takes precedence over the emoji; Remove resets it.
Emoji: a short symbol (up to 4 characters) shown on the brand colour.
Colour: the brand colour (hex #RRGGBB) for the icon background and card.
Name (required, max. 120 chars) and Description (max. 500 chars).

Behaviour

System prompt (required, max. 20,000 chars): the core instruction that defines the assistant's behaviour in every chat – role, tone, rules, boundaries.
Optimize (✨): lets the language model rewrite your prompt (sharper role, clear rules, same language). Undo restores the original.
Default model: the preferred model for chats with this assistant. — means the normal model picker applies.
Web search: lets the assistant research the web when needed (on by default).
Force web search on every answer (force_web_search, off by default): automatically runs a web search before every answer and feeds in the results. More reliable than leaving it to the model – but only sensible for pure research assistants.
Integrations: bind providers (e.g. Outlook, Google Drive, Jira) to the assistant. Their tools are enabled automatically in chats with this assistant. The user still needs their own OAuth connection to the provider.
Creativity (temperature, 0.0–1.0): how creative the model may be. Recommendation: 0.7 for assistants without attached knowledge, 0.3 with attached knowledge.
Advanced model parameters (collapsible): Max tokens (maximum answer length, 1–32,000) and Reasoning effort (low/medium/high, only relevant for reasoning models).

Input type: prompt or form

Prompt (standard): a normal chat – the user types freely.
Form: instead of a free-text chat, the user fills in predefined fields. Define them in the Fields section. Per field you set a key, label, description, whether it is required, and the type: text, longtext, select, number, email, checkbox, date or file (max. 20 fields).

Greeting message

An optional text (max. 2,000 chars) that the assistant automatically "says" as the first message of every new chat. Ideal for setting context or asking up front for the information you need.

Conversation starters

Up to four pre-written opening prompts (max. 500 chars each). They appear as clickable cards before the user types their first message – one click drops the text straight into the input.

Knowledge

Attach documents the assistant uses automatically when answering (RAG). There are two sources:

Link company-wide documents – references existing company-wide documents.
Own documents – uploaded directly for this assistant via Upload file or Pick from cloud (OneDrive, SharePoint, Google Drive).

In chats with this assistant, knowledge retrieval is restricted to exactly these documents (plus company-wide/project ones).

ℹ️

Knowledge vs. attachments: the knowledge stored here is permanently available in every chat with the assistant. It does not replace attaching individual files directly inside a chat.

Skills

Attach skills that are active only within this assistant. The assistant can call them itself via a tool when needed. Attach skill opens a searchable picker of the active skills.

Subagents

Link other assistants in your company as subagents to which this assistant can delegate sub-tasks. Each subagent works independently (its own knowledge, possibly web search) and returns its result.

⚠️

Delegation is deliberately bounded: at most two levels deep and with a call budget per answer. An assistant cannot be its own subagent, and direct cycles (A→B while B→A already exists) are rejected.

By default every company member can see and use the assistant. Here you grant specific people extended rights (by email address):

Role	Meaning
`Execute`	may select and use the assistant in chat
`Read`	may view the configuration but not chat with it
`Edit`	may modify the assistant (acts as a co-creator)

Visibility

Project scope: Company-wide (default) or a specific project. With a project scope the assistant is aimed at project members, and its knowledge retrieval additionally considers the project's documents.
Pin to sidebar: adds the assistant as a quick-access entry in the sidebar.

Labels

Add up to eight labels (max. 30 chars each). They help others find the assistant in the Library via the label bar.

Usage

A read-only overview: number of Messages, Tokens consumed, estimated Cost (in EUR) and Last used.

Verification (admins / Kasimir staff only)

Mark as verified: adds a verification badge (shield icon) to the card.
Highlight in library: makes the assistant appear at the very top of the list.

Draft and publishing

This is the most important concept when editing an assistant.

While editing you change a draft. Save stores the draft but does not change behaviour in running or new chats.
Chats always use the most recently published version of the behaviour. Only Publish applies the current state (it saves automatically first).
Whenever the draft and the published version differ, the header shows Unpublished changes.

Draft, save, publish: chats always use the published version.

⚠️

Which fields get "frozen"? Only behavioural fields are published (snapshotted): system prompt, default model, web search, forced web search, creativity/tokens/reasoning, conversation starters, input type and form fields. Display fields like name, emoji, brand colour and avatar are always live immediately – they don't need publishing.

Test and clone

▶ Test: saves the current state and immediately opens a fresh chat with the assistant – so you test exactly what's in the form.
Clone: creates a copy (name suffixed with "(Kopie)"). The configuration and linked company-wide documents are copied. The assistant's own file uploads and its avatar image are not copied.

Using an assistant in chat

Launch a new chat bound to the assistant either from the Library (click a card) or from a pinned entry in the sidebar.
On start, the greeting message and conversation starters appear first (if configured).
During the conversation the assistant's persona, preferred model, attached knowledge and bound integrations apply automatically.
You can also address an assistant per message via an @-mention – then only that assistant answers that specific message.

Limits and notes

Length limits: name 120, description 500, system prompt 20,000, greeting 2,000 chars; up to 4 conversation starters (500 each), 8 labels (30 each), 20 form fields.
Avatar images may be at most 2 MB.
Bound integrations enable the tools automatically but do not replace the user's own OAuth connection to the provider.
Delete and managing shares are reserved for the creator and admins/owners; verification is reserved for admins and Kasimir staff.
Deletion is permanent and cannot be undone.

BackChat Tools in Detail NextWorkflows