Koppla logo

AI Chat

Get context-aware writing assistance from an AI assistant that understands your note, research, and fact-check results.

The Chat tab in Notis provides a conversational AI assistant that understands the context of your current note. It reads your note content, research results, and fact-check findings to give relevant, informed responses about your writing.

Accessing the Chat Tab

Open the AI tools panel by clicking the message icon in the editor toolbar, or click the Chat tab if the panel is already open.

How Context Works

The Chat assistant is not a general-purpose chatbot — it's aware of what you're working on. When you send a message, the assistant considers:

  • Your note content — The text currently in the editor
  • Research results — Any searches you've run in the Research tab
  • Fact-check findings — Results from any fact checks you've performed
  • Note title — The title of the current note

This means the assistant can discuss your specific content, suggest improvements based on your research, and help address issues raised by fact-checking — all without you needing to copy-paste context.

The assistant is focused on mental health content. It's designed to help with clinical writing, evidence-based content, and therapeutic communication.

Starting a Conversation

When the Chat tab opens, starter prompts appear to help begin the conversation:

  • "Summarize this note" — Get a concise overview of your current draft
  • "How can I improve this?" — Receive feedback on structure, clarity, and content
  • "What's the evidence for this topic?" — Explore the research basis for your note's subject

Click any starter prompt to send it, or type a custom message in the input field.

Example Use Cases

Getting Writing Feedback

After drafting a note, ask the assistant for specific feedback:

  • "Is the opening paragraph engaging enough for a client audience?"
  • "Are there any sections that could be more concise?"
  • "Does the tone feel professional but approachable?"
  • "What's missing from this psychoeducation piece?"

Working with Research

After running a search in the Research tab, the findings become part of the chat context:

  • "How does the research I found support my main argument?"
  • "Can you suggest how to incorporate these findings into the introduction?"
  • "What are the key takeaways from the research for my note?"

Addressing Fact-Check Results

After running a fact check, the results are available to the assistant:

  • "How should I revise the misleading claims?"
  • "Can you suggest better wording for the flagged statistics?"
  • "What qualifiers should I add to the unsupported claims?"

Content Development

Use the assistant to help develop ideas or fill gaps:

  • "Suggest 3 practical coping strategies I could add to this section"
  • "What are common misconceptions about this topic I should address?"
  • "Help me write a conclusion that summarizes the key points"

Conversation Features

Message History

Conversations are preserved per note. When you switch between notes, each note retains its own chat history. Returning to a note restores the previous conversation, so you can pick up where you left off.

Formatted Responses

Assistant responses are rendered as formatted markdown with headings, lists, bold text, and other formatting. This makes longer responses easy to scan and reference.

Copying Responses

Each assistant message has a copy button that copies the response with formatting preserved. When pasted into the Notis editor, the formatting carries over — headings, lists, and emphasis are maintained.

Tips for Better Results

Be specific about what you need: "Suggest ways to make the CBT section more actionable for clients" works better than "improve this."

Reference sections directly: "The third paragraph about coping strategies needs more depth" gives the assistant clear direction.

Ask follow-up questions: The assistant remembers the conversation within a note session. Build on previous responses to refine suggestions.

Combine with other tools: Research a topic first, then ask the Chat to help incorporate findings. Or run a fact check, then ask for help revising flagged claims.