TotalApp Docs

AI Assistant Settings

Configure the embedded AI panels that appear on Kanban, HR Reports, Ecommerce, and other screens.

Overview

TotalApp's AI assistant panels are embedded directly into their host screens — they understand the context of the data on screen and can answer questions, generate reports, and perform analysis without you needing to copy-paste data into a separate tool. All assistant settings are global and apply to every panel across the app.

AI Assistant settings live in Settings → AI Assistant.

Panel Width

The assistant panel is a fixed overlay anchored to the right edge of the viewport. Three widths are available:

Small
288px
Medium
320px
Large
384px

When the panel is open, the host screen's content area reserves right-side padding equal to the panel width so that table columns and action buttons are not hidden behind the panel.

Auto-Open

When enabled, the AI assistant panel opens automatically when you navigate to a screen that has an embedded assistant. Useful if you rely on the assistant frequently. When disabled (default), you open the panel manually via the vertical tab anchored to the right viewport edge.

Typing Indicator

Shows an animated thinking spinner with a "Thinking..." label while the AI is generating a response. Toggle it off if you find it distracting. When off, the panel simply shows no new bubble until the response arrives.

Suggestion Chips

On an empty chat (no messages yet), the panel shows a list of example question chips below the glowing Sparkles icon. Clicking a chip sends it as your message immediately. Toggle off to start with a clean empty state — useful if you already know exactly what to ask.

Font Size

Controls the text size of chat bubbles in the assistant panel:

  • Small — 11px. Fits more text in the panel, especially useful at the Small panel width.
  • Medium — 12px (default). Standard readable size for extended conversations.

History Depth

Controls how many previous message pairs (user message + AI response) are included as context when you send a new message. Range: 1–20; default: 10.

  • Higher depth — the AI has more conversation context, leading to more coherent multi-turn conversations. Requests take longer and cost more tokens.
  • Lower depth — faster responses with less context. Best for single-shot analytical questions where conversation history is not needed.

Use New Chat for Fresh Context

Click the New Chat button (pencil square icon) in the assistant panel header to start a fresh session. This clears the current conversation from context (but saves it in the History panel), so subsequent questions are not influenced by previous exchanges.

Supported Screens

ScreenData ContextCan Save Reports?
Kanban BoardAll tasks in all columns across all projectsNo
My TasksUser's own tasks (all or selected project)No
My NotesUser's notes (all or active note)No
My ProjectsAll projects with member counts and statusNo
My WorkflowsSaved workflow graphs with node/edge detailsNo
My PresentationsSaved presentations with slide contentNo
HR ReportsEmployee, leave, expense, and overtime dataYes
Ecommerce ReportsLive Shopify products, orders, and customersYes
Course BuilderCurrent course snapshot (lessons, sections, tests)No
Shopify IntegrationLive Shopify Admin GraphQL dataNo

Frequently Asked Questions

Are AI Assistant settings per-screen or global?
All AI Assistant settings — panel width, auto-open, typing indicator, suggestion chips, font size, and history depth — are global. Changing any of them in Settings → AI Assistant instantly affects every embedded assistant panel across TotalApp (Kanban, HR Reports, Ecommerce Reports, My Tasks, My Notes, My Projects, My Workflows, My Presentations, Course Builder, Shopify Integration, and any future assistant-enabled screen). There is no per-screen override.
Why can some screens save reports and others can't?
Only screens designed around markdown report generation — currently HR Reports and Ecommerce Reports — expose a "Save as Report" action in the assistant panel's bottom dock. Screens like Kanban Board, My Tasks, or My Notes use the assistant purely for conversational Q&A over your data and don't produce a report artifact, so no save action appears for them.
Does increasing History Depth cost more or slow down responses?
Yes. History Depth controls how many previous message pairs are sent as context with each new request. A higher depth (up to 20) gives the AI more conversational memory but increases the token count of every request, which can make responses take longer and consume more of your usage quota. If you mostly ask one-off analytical questions, keep History Depth low (1–3) for the fastest responses.
How do I get a clean conversation without losing my previous chat?
Click the New Chat button (pencil-square icon) in the assistant panel header. This resets the active conversation's context to zero, but the previous conversation is not deleted — it is saved to the panel's History list, accessible via the Clock icon, so you can reopen it later without it influencing your new questions.