Understanding chat
Chat in Agentic studio is where you work with AI and manage your conversations. It provides a conversational workspace where you can create, review, and continue work while preserving full context and history.
The Chat page
The Chat page provides a centralized workspace for creating, reviewing, and continuing conversations. This includes chats started directly in the Chat interface, as well as conversations generated when running agents or flows.
To access it, in the navigation menu, click Agentic. Then, in the left sidebar, select Chat.
The page includes a History pane on the left, where you can browse and reopen previous conversations. Selecting a conversation loads it into the main workspace so you can review details, reuse outputs, or continue working.
On this page, you can:
-
Browse your chat history by clicking the
icon. -
Start a new chat by clicking the
icon or use suggested prompts. -
Open saved conversations to review messages and AI-generated artifacts.
-
Continue a conversation by adding context, tools, and uploading files.
-
Reuse AI-generated artifacts through agent sequencing within the same chat.
-
Convert a chat into a space for collaboration.
-
Delete chats that are no longer needed
Continue a conversation and reuse results
When you reopen a chat, all previous progress remains in the same thread. This allows you to:
-
Refine outputs with follow-up prompts or clarifications.
-
Ask additional questions to expand or validate results.
-
Reuse earlier responses and artifacts as context.
-
Run additional agents in sequence without restarting the task.
Artifacts generated by agents appear directly in the conversation and can be selected as reusable context for future steps.
Chat features and tools
Within any conversation, you can access a range of built-in chat features and tools. These are the same capabilities available in the Chat interface on the Overview page.
Type your question or request in natural language. You can also add more context, file attachments, or selected tools.
To submit a prompt, press Cmd + Enter (Mac) or Ctrl + Enter (Windows).
Examples:
-
Write 5 email subject lines for [Campaign]. Make one of them a little mysterious
-
Make this sound less like a corporate brochure and more like a text to a friend
-
Summarize this product update in three bullet points for executives
Choose a prebuilt or custom agent to handle your task. Depending on the agent, you may need to configure parameters and can optionally select items (pages, artifacts briefs, or CSV files) for the agent to act on.
Examples:
-
Select the Content Generator agent to draft marketing copy.
-
Select the Researcher to analyze emerging industry trends.
-
Select the Translator and select the target languages to localize the content.
Click Context 
to add sources that ground the AI. This helps it better understand your request and produce more relevant results. Context represents the information used to guide and inform the AI, giving you more control over how outputs are generated.
Available context sources:
Click Skills & Tools 
to give the AI additional capabilities.
Available capabilities:
Click Attach files 
to add one or more files related to your request. The AI can use attached briefs, documents, or drafts to generate more accurate and tailored responses.
Examples
Upload a product brief and ask the AI to draft a launch announcement.
Paste a blog draft and request edits for clarity and tone.
Click Voice input 
to speak your prompt instead of typing it.
Examples
Can you turn this into a quick social post?
Summarize this in three key points for my team.
Choose a Model Type to control how AI handles your request.
-
Default - automatically selects the best model for the task.
-
Fast - optimized for speed. Best for simple requests and lightweight tasks.
-
Expert - optimized for advanced reasoning and structured thinking. Responses may take longer.
Examples
Use Fast for quick rewrites or short summaries.
Use Expert for strategic planning, structured analysis, or complex problem-solving.
Agent sequencing within a single chat
Agent sequencing allows you to reuse the output of one agent as input for another agent within the same conversation.
For example:
-
The Researcher agent generates a Research Document for a specific topic. The document appears in the chat as an artifact.
-
When you reopen the conversation in Chat, you can select the existing artifact as context and run another agent.
-
You can select the research document and run another agent, such as the Bulk Content Generator, to build on the previous output.
The second agent uses the earlier results to complete the next task, allowing agents to run sequentially within the same conversation.
This allows you to:
-
Chain multiple agents within a single conversation.
-
Share context and outputs across agents.
-
continue evolving your work over time without restarting the chat.
Alternatively, you can chain agents for multi-step workflows by adding more agents within a space.
Convert a chat into a space
You can easily convert any conversation in Chat into a dedicated space. This allows you to continue working in a more structured environment, while preserving the full conversation history and associated artifacts. Once converted, the space can also be shared with other users for collaboration.
To convert a chat into a space:
-
In the navigation menu, click Agentic. Then, in the left sidebar, select Chat.
-
Start a new chat or open an existing one by selecting it from the history
. -
In the top left corner, click Convert to space.
A new space is created from the conversation and opens automatically, with all context and history preserved.
