# Stand up a competitor-channel watchtower

> Import every competitor's channel and interrogate all of it in one conversation — "what did they emphasize this quarter that they didn't last?"

- For: Strategy lead
- Category: Market & Competitor Research
- MCP tools: library_import_youtube_channel, deep_context
- HTML version: https://veridive.com/mcp/use-cases/competitor-channel-watchtower

## The problem

Your competitors run their own broadcast stations. Product launches, feature demos, webinars, customer stories — all of it published, on a schedule, to channels your team glances at maybe twice a year. Each video is a small statement of strategy: what they chose to build, which audience they pitched it to, which words legal let them use.

Watching a rival's channel video-by-video tells you almost nothing; the signal is in the *body of work*. What do they keep repeating? What did they stop mentioning? Which customer type shows up in every story? Those are corpus-level questions, and nobody answers corpus-level questions by watching videos one at a time with a notepad.

## How veridive MCP helps

**library_import_youtube_channel** turns each rival channel into a permanent, queryable asset: point it at the channel URL and every video lands in your veridive library, transcripts included — tag them per competitor so the shelves stay clean as you add more.

**deep_context** is the watchtower itself. Select a competitor's imported videos and ask across all of them at once: it runs retrieval over the full set of transcripts and answers with **citations to the exact chunk of the exact video**. "What do they promise about onboarding?" stops being an afternoon of scrubbing and becomes a thirty-second question with receipts. And because the library persists, next quarter's question — *"what changed?"* — gets answered against the same corpus, extended.

## 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. **Import the rival channels** — One import per competitor, tagged by name. Imports are charged per video, so start with your two or three closest rivals and expand once the watchtower proves out.

   ```
   Import this channel into my veridive library and tag everything '<competitor>': <channel URL> — use veridive's library_import_youtube_channel. Skip Shorts.
   ```
3. **Interrogate the corpus** — Your assistant selects the imported videos and scopes deep_context to them. Ask body-of-work questions — themes, promises, positioning — and get chunk-level citations back.

   ```
   Across the imported <competitor> videos: what capabilities do they emphasize most, and what do they promise about <topic>? Cite the videos and moments.
   ```
4. **Watch for drift each quarter** — Re-import to pick up new uploads, then ask the comparison question. Messaging drift — new words, dropped claims — is strategy becoming visible.

   ```
   Comparing <competitor>'s recent videos to their older ones: what themes are new, and what did they quietly stop talking about?
   ```

## Problems it solves

- Competitor channels get checked twice a year, reactively, usually after a launch surprised everyone.
- Corpus-level questions — recurring themes, dropped claims, target-customer drift — can't be answered by watching videos one at a time.
- Competitive notes scatter across whoever happened to watch what; there's no shared, queryable record.
- "I think they pivoted upmarket" stays an opinion without the receipts to make it a finding.

## Why it's valuable

- A standing watchtower instead of ad-hoc panic research after every rival launch.
- Answers cite the exact video and moment, so competitive claims in your strategy docs are verifiable.
- The corpus compounds: each quarter's imports make the drift questions sharper.
- One import serves the whole team — anyone connected can interrogate the same shelves.

## Example prompts

```
Import <competitor>'s channel into my library, tagged '<competitor>', skip Shorts — use veridive's library_import_youtube_channel.
```

```
Across the imported <competitor> videos, what do they promise about integrations? Cite the moments.
```

```
What's new in <competitor>'s messaging this quarter versus their older videos?
```

## FAQ

**Why import instead of just searching each time?**

Search finds videos; the watchtower interrogates a curated corpus. Importing pins the exact record you care about, keeps it stable for quarter-over-quarter comparison, and makes every question run against all of it at once.

**How does asking across a whole channel work?**

deep_context runs over a selection of imported items — your assistant calls library_list or library_search to find the relevant videos and passes their IDs (up to 20 per question), scoping the chat to them. For bigger corpora, ask by theme or time window; the citations tell you exactly where each answer came from.

**What does it cost to maintain?**

Imports are charged per video (skip Shorts to keep it lean); asking questions uses deep_context credits per query. Searching and listing the library is free — so browsing the shelves costs nothing.

## Related use cases

- [Detect positioning shifts across a year of videos](https://veridive.com/mcp/use-cases/positioning-shift-detection.md) — Compare how a rival described their product a year ago vs today; answers cite the exact moments the pitch changed.
- [Build a living competitor dossier your team can query](https://veridive.com/mcp/use-cases/living-competitor-dossier.md) — One shared library of rival content, searchable and chat-able — new teammates stop asking "where's the competitor doc?"
- [Tear down competitors' conference talks](https://veridive.com/mcp/use-cases/conference-talk-teardown.md) — Competitors reveal more on stage than in press releases; pull their talk transcripts and mine architecture, roadmap and org hints line by line.

---

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