VDF AI Data

VDF AI Data

The layer that turns your files, folders, and connected apps into reliable, searchable knowledge VDF AI can use to answer questions and produce grounded outputs.

3 min read
On this page

The quiet layer that makes everything else better

You can have a great agent and a well-designed network — and still get disappointing results if the underlying sources are wrong, stale, or invisible. VDF AI Data is the layer that fixes that.

Think of it as the bridge between the documents your team already produces — proposals, reports, notes, transcripts, specs — and the questions you ask VDF AI every day. When sources are clean, current, and findable, every chat, every agent, every network suddenly gets sharper.

Most teams underestimate this layer. The teams that get the most value out of VDF AI spend disproportionate time setting up Data well — and reap that investment across every product surface.

Who VDF AI Data is for

Anyone whose work depends on real source material — files, folders, connected apps — being reliably usable by the AI.

  • Knowledge workers
  • Research analysts
  • Operations leads
  • Customer success
  • Product teams
  • Legal & compliance
  • Sales engineering
  • Workspace admins

A few signals you’ll spend a lot of time in Data:

  • Your day-to-day work depends on internal documents — policies, reports, customer notes.
  • Your team has scattered sources across Google Drive, Confluence, GitHub, Box, SharePoint.
  • You want answers that cite something, not invent something.
  • You generate the same kinds of deliverables from large source documents — turning a 60-page contract into a checklist, or a quarter of standup notes into a summary.

What you can do with VDF AI Data

Upload files for use anywhere

Drop in PDFs, text files, spreadsheets, slides, and transcripts. Reuse them in any conversation, agent, or network.

Connect the apps your team lives in

Google, Microsoft, Confluence, Jira, GitHub, Slack, Zoom, GitBook — all referenceable inside the AI without copy-pasting.

Ask questions across your knowledge

Search by meaning rather than keyword — "find every reference to renewal terms" surfaces the right paragraphs even if the exact word "renewal" isn't there.

Produce grounded outputs

Turn source material into summaries, checklists, comparisons, and structured documents with citations back to the source.

Compare versions and find changes

Diff two policy drafts. Highlight what changed between revisions. Trace decisions back to specific edits.

Keep sources current

Refresh connections, swap out stale folders, and prune sources that are no longer useful — without rebuilding everything from scratch.

How VDF AI Data differs from “just uploading files to chat”

You can drop a file into Chat any time. So why does Data exist?

FeatureDrop a file into ChatVDF AI Data
One-off use✅✅
Reuse across conversations—✅
Reuse across agents—✅
Reuse across networks—✅
Searchable across all sources—✅
Stays current as sources update—✅
Shared with the team—✅
Citations linking back to sourcepartial✅

The simple rule: if you’ll only use this file once, drop it in Chat. If it’ll be used more than once — by you or anyone else on the team — bring it into Data.

Key concepts, in plain language

  1. Source.

    Anything VDF AI can read on your behalf — an uploaded file, a connected app folder, a Confluence space, a Jira project, a Slack channel.

  2. Connection.

    The link between VDF AI and a connected app. It controls what VDF AI can access and stays current as the app changes.

  3. Search by meaning.

    The ability to find content based on what you mean, not just the words you typed. "Renewal terms" will surface paragraphs about contract extensions even if "renewal" isn't in them.

  4. Grounded output.

    An AI-produced result that ties its claims back to specific sources you can verify.

  5. Workspace knowledge.

    The total set of sources your workspace can draw on. Some sources are personal, some are team-shared, some are workspace-wide.

  6. Refresh.

    Updating a connection so VDF AI sees the latest state of an app. Connections can refresh automatically or on demand.

A quick mental tour

Three quick ways VDF AI Data shows up in real work:

1. You upload a contract and ask for a checklist

You drop a 40-page MSA into Data. You open Chat or an Agent and ask, “Turn this MSA into a checklist of obligations for our delivery team.” The output is a clean checklist that cites which page each obligation came from.

2. You connect Confluence and ask across teams

You connect your team’s Confluence space. From any conversation, you can ask, “What does our team say about onboarding new customers?” — and VDF AI pulls from the relevant Confluence pages, returning an answer with links back to each page it referenced.

3. You compare two versions of a doc

You upload last quarter’s policy and this quarter’s draft. You ask, “What changed?” — and VDF AI returns a section-by-section list of the substantive differences, citing where each one appears.

In all three cases, Data is doing the invisible work: making the source material readable, searchable, and reliably reusable.

How VDF AI Data fits with the rest of the platform

  • VDF AI Chat uses Data sources implicitly whenever you reference files or connected apps.
  • VDF AI Agents pull from Data sources to ground their outputs and produce verifiable claims.
  • VDF AI Networks use Data sources at one or several stages — research stages especially benefit.

Investing in Data once pays back across all three other products.