​

Workspace Chat

The Workspace Chat is your AI-powered strategy partner. Unlike basic AI chat, it knows your brand guidelines, target audiences, content inventory, and performance data. Attach skills to guide the AI through specific workflows.

​

Getting Started

Open Chat from the main navigation. You’ll see a full-page chat interface with:

• Conversation sidebar (left) — your chat history with search, pin, and archive
• Message area (center) — the conversation with markdown rendering and inline artifacts
• Input bar (bottom) — message input with skill chips and model selector

​

Sending Messages

Type a message and press Enter to send (Shift+Enter for a new line). The AI responds with streaming text, and can execute tools to look up your data during the conversation.

​

Suggestion Chips

When starting a new conversation, you’ll see suggestion chips to get started:

• “What are my top performing content topics?”
• “Create a content calendar for next month”
• “Analyze my SEO opportunities”
• “Review my brand messaging”

Click any chip to send it as your first message.

​

Editing Messages

You can edit any message you’ve sent to refine your request or take the conversation in a different direction. To edit a message:

1. Hover over your message to reveal the pencil icon on the left
2. Click the pencil to enter edit mode
3. Modify the text in the textarea that appears
4. Click Save & Send to submit the edited message, or Cancel (or press Escape) to discard changes

When you edit a message, the conversation forks — everything after your edited message is replaced with a fresh response based on your revised text. The original conversation is preserved as a branch you can return to.

​

Conversation Branches

After editing a message, a small branch navigator appears below it showing something like < 1/2 >. Use the arrows to switch between different versions of the conversation from that point forward.

• Switching branches is instant and free — no AI calls are made
• The most recent branch is always loaded when you return to the conversation
• Each edit creates a new branch, preserving your conversation history

​

Skills

Skills are reusable AI playbooks that guide the AI through specific workflows. See Skills for details on creating them.

​

Attaching Skills

1. Click + Skill in the input bar (only visible with Claude models)
2. Browse skills by category or search by name
3. Click a skill to attach it — a chip appears in the input bar
4. If the skill requires inputs (e.g., target keyword, time period), fill out the inline form
5. Send your message — the AI receives the skill instructions plus all resolved data

​

Multiple Skills

You can attach multiple skills at once. Each skill’s instructions and data are included in the AI’s context. This is useful for complex workflows that span multiple domains.

​

Detaching Skills

Click the x on a skill chip to remove it. The AI will no longer follow that skill’s instructions in subsequent messages.

​

Model Selection

Choose your AI model from the dropdown in the input bar:

​

Conversations

​

Managing Conversations

• New Chat — Click “New Chat” in the sidebar to start fresh
• Search — Filter conversations by title in the sidebar search
• Pin — Pin important conversations to the top of the list
• Rename — Click the menu on any conversation to rename it
• Archive — Hide old conversations (recoverable)

​

Auto-Titles

After your first exchange, the AI automatically generates a short title for the conversation (e.g., “SEO Content Calendar” or “Q2 Campaign Ideas”). Titles appear in the sidebar for easy navigation.

​

Conversation History

Conversations are grouped by date: Today, Yesterday, Previous 7 Days, and Older. Pinned conversations always appear at the top.

​

Goal-Driven Mode

When you open a work item from the Inbox via the “Work on this in chat” button, the chat enters goal-driven mode. A Goal Bar appears above the conversation showing progress through the asset’s lifecycle phases.

​

How It Works

The Goal Bar displays a horizontal progress tracker with the phases required to complete the asset — typically Setup, Outline, Draft, and Publish, though the exact steps vary by asset type. Each step shows a status icon:

• Green check — completed
• Blue pulse — currently active
• Gray circle — pending

As you work through the conversation, steps auto-advance when the AI detects you have completed the current phase. For example, once you confirm an outline, the bar moves from Outline to Draft automatically. You do not need to manually advance steps.

​

Automatic Skill Attachment

In goal-driven mode, skills are automatically attached based on the current lifecycle step. When the bar advances, the attached skill updates to match the new phase — no manual skill management needed.

​

Supported Asset Types

Goal-driven mode works for all 9 asset types: blog posts, commercial pages, ad creatives, CTAs, nurture sequences, social content, agent prompts, video scripts, and sales enablement content.

​

Inline Artifacts

When the AI produces structured output, it renders as interactive cards inline in the conversation: Artifacts can be collapsed/expanded when they’re long.

​

Tools

The AI has access to tools that let it look up your data during the conversation:

• View ICPs — See your target audience profiles
• View Hooks/Angles — Browse your messaging library
• Search Content — Find existing content by keyword
• Generate Ideas — Create content ideas
• Check Brand Consistency — Validate against guidelines
• Find Content Gaps — Identify missing coverage

Tool execution appears as status indicators in the message stream (spinner while running, checkmark when complete).

​

What the AI Knows

The workspace chat automatically includes:

• Your brand position and core messaging
• Your editorial guidelines
• Your target audience profiles (ICPs)
• Your confirmed learnings from past experiments
• Relevant content from semantic search (RAG)
• Skill instructions and data (when skills attached)

You don’t need to repeat context — the AI already has it.

​

Access Control