# Build an interview dossier on any public figure

> Every appearance, the stories they always tell, the questions they dodge — ask what no one has.

- For: Producer / Host
- Category: Media, PR & Comms
- MCP tools: deep_search, deep_query, run_youtube_tool
- HTML version: https://veridive.com/mcp/use-cases/public-figure-interview-dossier

## The problem

Your guest has done this before. Whoever sits across from you next week has given the same interview forty times: the same origin story polished to a shine, the same three anecdotes deployed on cue, the same graceful pivot away from the one topic they'd rather not touch. Most hosts walk in having skimmed two of those forty appearances — and deliver interview forty-one.

The great interviews happen when the host has heard all forty: knows which stories are pre-loaded (and skips them), notices what the guest has *never* been asked, and catches the thread from an obscure podcast two years ago that deserves pulling. That's not talent; that's preparation at a volume nobody has time for. Which is exactly the kind of preparation you can now delegate.

## How veridive MCP helps

**deep_search** assembles the full appearance record — interviews, podcasts, panels, sorted by date, the obscure alongside the famous. **run_youtube_tool**'s `key-quotes` and `podcast-notes` compress each appearance into its beats: the stories told, the claims made, the moments that got clipped.

**deep_query** then does the dossier's real work — the cross-appearance read: *"which anecdotes does this person repeat verbatim?"* (skip those), *"what topics do they consistently deflect?"* (decide whether to press), *"what did they say years ago that contradicts their current line?"* (the question nobody else will have). Every finding cited to its moment, so your producer can verify the ammunition before you fire it. You walk in as the forty-first interviewer who somehow heard the first forty.

## How to run it

1. **Connect veridive MCP** — One-time setup: add the veridive MCP server to Claude, ChatGPT, Cursor or any MCP client and generate your API key. Takes about two minutes.
2. **Pull the full appearance record** — Everything, not just the greatest hits — the obscure appearances are where the unrehearsed material lives.

   ```
   Find all interviews, podcasts and panel appearances with <guest>, across all years, including smaller channels — use veridive's deep_search.
   ```
3. **Compress each appearance to its beats** — Key quotes and notes per appearance: the stories, the claims, the recurring lines.

   ```
   For these appearances, pull key quotes and podcast notes — I want each one's main stories and claims: <links>
   ```
4. **Run the cross-appearance read** — The dossier questions: what's rehearsed, what's dodged, what's contradictory, what's never been asked.

   ```
   Across <guest>'s appearances: which stories do they repeat, which topics do they deflect, and where do their past statements contradict their current line? Cited moments for each.
   ```

## Problems it solves

- Guests arrive with forty rehearsed interviews behind them; hosts arrive having skimmed two.
- The unrehearsed material hides in obscure appearances no prep sheet ever covers.
- Rehearsed-anecdote detection requires cross-referencing every telling — impossible manually.
- The question nobody has asked is invisible until you know everything that has been.

## Why it's valuable

- Prep depth that used to take a research team, delivered before one production meeting.
- Every dossier claim cites its moment — producers verify before air, not after.
- Skipping the pre-loaded stories buys minutes of interview time back for the material nobody's heard.
- The contradiction thread and the never-asked question are what make interviews get clipped — yours, this time.

## Example prompts

```
All appearances by <guest>, all years, including small channels — use veridive's deep_search.
```

```
Key quotes + podcast notes for these appearances: <links>
```

```
Across <guest>'s record: repeated stories, deflected topics, contradictions — cited.
```

## FAQ

**How long does a full dossier take?**

The sweep and compression run in minutes; the cross-appearance read is a few prompts more. Budget an hour for a dossier that used to be a researcher's week — most of it your reading time, not the machine's.

**Does this work for panel moderation and event prep too?**

Anywhere you interview someone on the record: podcast guests, panelists, keynote Q&As, even executive candidates who speak publicly. The rehearsed-versus-fresh distinction is universal.

**Isn't using contradictions adversarial?**

That's an editorial call — the dossier just makes the record visible. Plenty of hosts use the same thread generously: 'you used to say X, what changed?' is often the question guests most enjoy answering.

## Related use cases

- [Prep for a podcast guest in 30 minutes](https://veridive.com/mcp/use-cases/podcast-guest-prep.md) — Every prior appearance, key quotes and podcast notes — walk in knowing which stories they always tell, and ask what nobody has.
- [Hunt sources with timestamped quotes](https://veridive.com/mcp/use-cases/timestamped-source-hunting.md) — Find who said it, pull the transcript, quote to the second — attribution your editor can verify in one click.
- [Analyze how your spokesperson actually lands](https://veridive.com/mcp/use-cases/spokesperson-message-analysis.md) — Import their appearances, extract key quotes, and check delivered-message vs intended-message drift.

---

Part of the veridive MCP use-case library: https://veridive.com/mcp/use-cases.md
Server docs for agents: https://veridive.com/mcp.md