MCP

AI-agent format

Steal your customer's exact words

Mine how real users describe the problem in reviews and vlogs — copy that converts is usually a transcript quote.

Product marketerSales, Marketing & Branddeep_searchlibrary_import_youtube_videoget_youtube_transcript
The use case

The problem

Your landing page says "streamline your workflow with intelligent automation." Your customer, describing the same problem to a friend, says "I spend every Friday copying numbers between three tools and I want to die." One of these phrasings converts; you paid an agency for the other one.

Voice-of-customer copy is the oldest law in marketing — write in the words buyers use, not the words vendors invent — and the reason everyone breaks it is supply: getting real customer phrasing means interviews, surveys, review-mining. Meanwhile people describe their problems on camera constantly: review videos, day-in-the-life vlogs, "how I finally fixed X" tutorials. The phrasing goldmine is public and spoken; it just needs mining.

The mechanics

How veridive MCP helps

deep_search finds where your buyers talk: reviews of your category, workflow vlogs, complaint-driven tutorials — transcript-level search, so "videos where people describe struggling with <problem>" actually returns those moments, evidence-timestamped.

library_import_youtube_video and get_youtube_transcript extract the ore: import the richest videos and pull the verbatim passages where real users describe the pain, the workaround and what "fixed" feels like — in the exact vocabulary your headlines should be wearing. The phrasing file that comes out (frustration verbs, before/after language, the words they don't use — usually your feature names) is the cheapest conversion-rate work you'll do this quarter.

Step by step

How to run it

  1. 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. 2
    Find where buyers describe the pain

    Reviews, vlogs and tutorials where real users narrate the problem your product solves.

    Find videos where real users describe struggling with <problem your product solves> — reviews, vlogs, honest tutorials, not vendor content — use veridive's deep_search.
  3. 3
    Mine the verbatim phrasing

    Import the richest ones and pull the passages — the exact words, complaints and metaphors.

    Import these and pull the transcript passages where they describe the problem and what fixing it felt like — verbatim: <video URLs>
  4. 4
    Build the phrasing file

    Recurring vocabulary, organized by funnel stage — the raw material for every headline test.

    From these passages: the recurring words and phrases for the pain, the workaround, and the relief — grouped, with which video each came from.
Before vs after

Problems it solves

  • Copy gets written in vendor vocabulary because customer phrasing is expensive to collect.
  • Interviews and surveys yield sanitized language; the raw phrasing comes out when users talk to cameras, not researchers.
  • Agencies invent 'streamline' and 'seamless' while buyers say 'I want to stop doing this by hand'.
  • Feature names your team loves appear nowhere in how customers describe the problem — and nobody checks.
The payoff

Why it's valuable

  • Headlines built from buyers' own words — the conversion lever with a century of evidence behind it.
  • The phrasing arrives verbatim and sourced, ready for copy tests, sales decks and ad variants.
  • The negative finding matters too: the vocabulary gap between your site and your buyers, made visible.
  • Refreshable per segment and per quarter — voice-of-customer stops being a one-off research project.
Copy & paste

Example prompts

Videos where real users describe struggling with <problem> — no vendor content — use veridive's deep_search.
Import these; pull verbatim passages on the pain and the relief: <URLs>
Phrasing file: recurring pain/workaround/relief language, grouped and sourced.
Questions

FAQ

Why spoken phrasing over review-site text?

Speech is less performed — G2 reviews are written for an audience with a template; someone narrating their workflow on camera reaches for the words they actually think in. Both are useful; the spoken layer is the one nobody mines.

Which tools does the mining chain?

deep_search to find the moments (transcript-level), then library_import_youtube_video + get_youtube_transcript for verbatim passages — pulls are free once a video is imported.

How do I use the file beyond headlines?

Everywhere language touches conversion: ad copy, onboarding text, sales-call vocabulary, even roadmap naming. The file is a shared asset — PMM builds it, everyone borrows from it.

Keep going

Related use cases

Run this in your own assistant

Connect the veridive MCP server to Claude, ChatGPT or Cursor and paste the prompts above — setup takes about two minutes.