VDF AI Agents FAQ

Choosing the right agent

I don’t know which agent to choose. Where do I start?

Pick by the output you want, not the topic you want to discuss. If you want to ship a release note, look for a release-note drafter. If you want a meeting summary, look for a summary specialist.

If two agents seem to fit, pick the narrower one — it’ll make fewer assumptions and produce more consistent results.

Still stuck? Use VDF AI Chat in agent-discovery mode and describe what you’re trying to produce. It’ll suggest a starting point.

Can I see what an agent is good at before I use it?

Yes — every agent has a short description that names its outcome, its preferred sources, and the kinds of audiences it works well for. Read the description before your first run. A 15-second skim usually points you at the right one.

My teammate can see an agent I can’t. Why?

A few possibilities:

• Workspace settings — some agents are scoped to certain workspaces. Ask your workspace admin.
• Plan level — some agents are part of higher tiers.
• Permissions — some agents need access to a connected app you may not have permission for.
• Recent additions — newer agents may not yet be rolled out to all users in your org.

Your workspace admin can check all four in one place.

Briefing and source quality

My output is close but the format isn’t right. What’s the fastest fix?

State the format explicitly in your refinement. Don’t ask for “better structure” — name the structure:

• “Use a three-bullet list, no preamble.”
• “Convert this into a Given/When/Then format.”
• “Make this a single paragraph, no headings.”
• “Format as a table with three columns: Item, Owner, Date.”

The agent didn’t use the file I attached. What happened?

Two common causes:

1. The agent doesn’t know which file to use. Name the file in your prompt: “Use the attached customer-call-2025-03-12.pdf, not the older transcript.”
2. The file format wasn’t readable. Scanned PDFs, very large CSVs, and password-protected docs can fail to parse. Try re-exporting as plain text or a smaller file.

The agent invented a fact. Why?

Agents can occasionally produce confident-sounding details that aren’t in your sources. Two things help:

• Be explicit: “Only use facts that appear in the attached transcript. If something isn’t there, say so.”
• Verify before you ship. Especially anything that names a customer, a number, a date, or a person.

My brief was long but the output was generic. Why?

Long briefs aren’t the same as specific briefs. A common pattern: the brief lists a lot of background but doesn’t name a clear outcome, audience, or constraints. Try cutting the background by 70% and adding:

• A one-sentence outcome.
• A one-sentence audience.
• Two or three hard constraints.

You’ll usually see a sharper output.

Choosing between Chat, Agents, and Networks

When should I use Chat instead of an Agent?

Use VDF AI Chat when:

• You’re exploring a problem and don’t know the shape of the output yet.
• The task crosses several different kinds of work (research + drafting + analysis).
• You’re not sure which agent to pick — Chat in agent-discovery mode will recommend one.

When should I use an Agent instead of Chat?

Use an agent when:

• You know the shape of the output you want.
• The task repeats across the team.
• You want predictable format on the first try.

When should I use Networks instead of an Agent?

Use VDF AI Networks when:

• The task has obvious stages (research → draft → critique → final).
• Several specialists need to combine for a single result.
• The workflow itself is the thing you want to standardize, not just the brief.

A good signal: if you find yourself chaining two or three agents by hand every week for the same task, that’s a network waiting to be built.

What about VDF AI Data?

Use VDF AI Data when the result depends mostly on your source material — files, folders, connected apps — and less on the brief itself. Data is the layer underneath that makes those sources usable. You’ll often use Data and Agents together: Data makes the sources clean, Agents turn them into deliverables.

Sharing and team usage

Can I save a successful brief for my team?

Yes. When you produce an output you love, save the brief as a template. Most workspaces let you mark it for yourself or for the team. The next person who needs the same kind of output starts from your win.

How do I find briefs my team has already saved?

Open the saved briefs area in your workspace (sometimes called Templates, sometimes Saved Prompts). Filter by team. If your workspace admin has set up team libraries, you’ll see the team’s collection there.

If you don’t see one yet, start one — a single shared brief is enough to plant the habit.

Can two people run the same agent at the same time?

Yes. Agent runs are per-user and per-conversation. Two people using the same agent at the same time get their own independent runs.

Where can I see my run history?

Your workspace keeps a history of recent conversations and agent runs. Open the activity area to scroll back. You can return to any previous run and continue refining — the context is preserved.

Privacy, sources, and trust

Where does the agent’s output come from?

Agent outputs are produced from your brief and the sources you provide. If you ask the agent to cite a source, it’ll point you to the file, document, or connected reference it used.

For deeper questions on data handling, see Privacy & Security.

Are my files shared with other users?

No. Files you upload and connected-app references are scoped to your workspace and the people you choose to share them with. An agent in your workspace doesn’t read another workspace’s files.

Can I delete a conversation or run?

Yes. You can delete individual conversations and runs from your activity area. Deletions remove the run from your history and from VDF AI’s working memory.

Still stuck?

• Working with assistants — the brief-context-output loop in detail.
• Tips & best practices — patterns that compound across your team.
• Talk to us — if your question isn’t here, we want to hear it.