VDF AI Documentation

VDF AI Chat

Use VDF AI Chat for conversations, document work, voice input, connected-app workflows, and shareable outputs.

What VDF AI Chat is for

VDF AI Chat is the main workspace for most daily product use. It combines conversational AI, file handling, guided workflows, and connected-app context in one place.

Use it when you want to:

  • ask questions in natural language
  • summarize or transform uploaded documents
  • draft stories, plans, briefs, or action lists
  • work with connected tools such as Jira, Google Drive, Slack, or Zoom
  • move quickly between open-ended questions and focused workflows

Two ways to work

Agent mode

Agent mode is best when you want VDF AI to decide the best path for you. You describe the goal and the platform chooses the most relevant assistant or workflow.

Good examples:

  • “Review this document and turn it into a rollout checklist.”
  • “Summarize what changed in our backlog and suggest next priorities.”
  • “Prepare talking points for tomorrow’s meeting using the notes in Drive.”

Manual mode

Manual mode is best when you already know the kind of outcome you want. Instead of asking VDF AI to choose the path, you pick the workflow directly.

Good examples:

  • story creation
  • backlog refinement
  • report analysis
  • other focused tasks available in your workspace

Core parts of the chat interface

  • Conversation area for your prompts and responses
  • Composer for typed prompts, pasted text, and attachments
  • Mode selector to switch between Agent mode and a focused workflow
  • Attachment controls for documents and other source material
  • Integration shortcuts to bring in connected data where available
  • Preview/export actions for generated outputs

Documents and attachments

VDF AI Chat becomes more useful when you give it the source material it should use. Depending on your workspace, you can work with:

  • uploaded files from your device
  • connected files from services such as Google Drive or Microsoft tools
  • pasted text or copied notes
  • context from documentation tools such as Confluence or GitBook

Good document-based prompts include:

  • “Summarize this file in five bullets.”
  • “Turn this specification into acceptance criteria.”
  • “Compare these two documents and highlight the differences.”
  • “Extract all action items and group them by owner.”

Voice input

If voice input is enabled in your workspace, you can dictate prompts instead of typing them.

Voice input is most useful when:

  • you want to capture longer ideas quickly
  • you are reviewing a document and speaking instructions out loud
  • you want a faster first draft before editing the prompt

After dictation, review the transcribed text before sending it.

Outputs you can expect

VDF AI Chat can return:

  • summaries
  • structured plans
  • user stories
  • acceptance criteria
  • checklists
  • meeting notes
  • comparison tables
  • drafts for follow-up work

Many responses can be copied, refined, exported, or reused as the starting point for another task.

Chat becomes stronger when you connect the tools that already hold your team context.

Tips for better results

  • Start with the outcome you want, not just the topic.
  • Mention the source you want VDF AI to rely on.
  • Attach the file instead of describing it from memory.
  • Use Agent mode when the task feels broad or multi-step.
  • Use a focused workflow when you want a repeatable output format.

Common issues

The answer feels too generic

Give more context. Mention the product, team, file, meeting, or system the answer should be based on.

I do not see the option I expected

Some workflows only appear when the right plan, permissions, or connected app is available in your workspace.

A connected action is missing

Reconnect the relevant app and refresh the page. If the connection expired, VDF AI will not be able to use that source until it is authorized again.

A document-based answer looks incomplete

Make sure the file uploaded successfully or that the connected app contains the document you expected.