# Build support FAQs from how-to videos

> The questions tutorials answer are the tickets you'll receive; draft deflection docs from them before launch.

- For: Support lead
- Category: Product & UX Research
- MCP tools: library_import_youtube_video, run_youtube_tool
- HTML version: https://veridive.com/mcp/use-cases/faq-from-howto-videos

## The problem

Support content has a timing problem: the FAQ gets written from the ticket queue, which means the queue has to fill before the deflection exists. Every recurring ticket answered manually while the docs catch up is cost that better timing would have deleted.

The tickets are forecastable. How-to videos about your product — official and third-party — are a map of what users need explained: every tutorial exists because a question recurs, every chapter marks a sub-question, every comment under it asks the follow-up. The ecosystem has effectively pre-written your FAQ's table of contents; support just needs to read it before the queue does.

## How veridive MCP helps

**library_import_youtube_video** collects the forecast corpus: the how-to and troubleshooting videos about your product, imported with transcripts. **run_youtube_tool** processes them into support assets: `how-to-extractor` pulls the procedural answer from each video into step form; `document` drafts the article; `comment-summarizer` on the same videos surfaces the follow-up questions the video *didn't* answer — the second layer of your future queue.

The output is a deflection library drafted from demonstrated demand: articles for the questions tutorials already prove people ask, written before those questions become tickets. Post-launch, the same pass on first-look videos (see the launch-triage use case) keeps the FAQ ahead of the queue instead of behind it.

## 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. **Collect the question map** — How-to and troubleshooting videos about your product — each one a proven recurring question.

   ```
   Import these how-to videos about <product> into my library, tagged 'support-map': <video URLs> — use veridive's library_import_youtube_video.
   ```
3. **Draft the deflection articles** — Procedural answers extracted per video, drafted into article form.

   ```
   For each imported 'support-map' video: extract the how-to steps and draft a support article answering the question the video answers.
   ```
4. **Mine the follow-up layer** — The comments under how-tos hold the questions the videos didn't answer — your queue's second wave.

   ```
   Summarize the comments on these how-to videos: what follow-up questions recur that the videos leave unanswered? <video URLs>
   ```

## Problems it solves

- FAQs get written from ticket queues — deflection arrives after the cost, not before.
- The ecosystem's how-to layer forecasts the queue and support never reads it.
- Procedural answers exist on video where search and ticket macros can't reach them.
- Follow-up questions under tutorials — the queue's second wave — go unmined entirely.

## Why it's valuable

- Deflection docs exist before the tickets do — the timing problem inverts.
- Every article maps to demonstrated demand: a tutorial someone bothered to make.
- The comment layer forecasts the second wave while it's still cheap to answer.
- Third-party how-tos double as a docs-gap audit: what outsiders explain, your docs didn't.

## Example prompts

```
Import these how-tos about <product>, tagged 'support-map': <URLs> — use veridive's library_import_youtube_video.
```

```
Per video: extract steps and draft the support article.
```

```
What follow-ups recur in these videos' comments, unanswered? <URLs>
```

## FAQ

**Which tools do the conversion?**

library_import_youtube_video for the corpus, then run_youtube_tool's how-to-extractor for step extraction, document for article drafts, and comment-summarizer for the follow-up layer.

**Our official videos cover these topics — why draft articles too?**

Different retrieval: tickets deflect through searchable text, not videos users won't watch mid-frustration. The article is the deflection surface; the video proves the demand.

**How does this stay current post-launch?**

New tutorials and first-look videos feed the same pipeline — a monthly pass keeps the deflection library tracking what the ecosystem newly explains.

## Related use cases

- [Treat tutorials as free usability tests](https://veridive.com/mcp/use-cases/tutorials-as-usability-tests.md) — Every workaround a tutorial teaches is a usability bug report; collect and query them across your product's ecosystem.
- [Triage creator first-looks after your launch](https://veridive.com/mcp/use-cases/first-look-feedback-triage.md) — Fresh-upload filters plus key-quotes compress launch-week reaction videos into a feedback sheet by end of day.
- [Repurpose webinars into SEO blog posts](https://veridive.com/mcp/use-cases/webinar-to-seo-posts.md) — That hour-long webinar is three blog posts; import it and draft them with structure and quotes intact.

---

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