Support Deflection Gap-Finder

SKILL: Support Deflection Gap-Finder
Every repeated "how do I" with no good doc, video, or search answer is a ticket you'll keep paying for. This skill mines high-frequency help-seeking across forums, Reddit and X, then cross-checks coverage on YouTube and the open web (including Google's AI overview) to expose where a single piece of content would deflect the most support volume. Use the optional platforms input to choose which surfaces run, and tune the demand window, region, and result depth.

You are running this skill on API Direct via its MCP tools. Execute the steps below yourself by calling the named tools in order — values in <angle brackets> come from a previous step. Then deliver the result described at the end.

INPUTS:
  - product (required): <product — ASK THE USER>
      Product whose help-content gaps you want to find
  - platforms (optional): (optional — e.g. forums, reddit, twitter, youtube, web)
      Comma-separated list of platforms to run — options: forums, reddit, twitter, youtube, web. Omit to run all of them; name specific platforms to limit the run.
  - demand_window (optional): year (default — override if you like)
      Recency window for help-seeking demand and web coverage checks (maps to the forums and web time param). Defaults to year.
  - region (optional): (optional — e.g. us)
      Optional country code to localize demand and coverage for region-specific products (maps to forums and web country). Leave blank for global.
  - depth (optional): 3 (default — override if you like)
      How many result pages to scan per surface (maps to the pages param on reddit/twitter/youtube). Higher = more questions and tutorials sampled.

PLATFORM SELECTION: some steps are gated with "Only run if {platforms} includes X". If the user supplied a `platforms` value, run only the steps whose platform is listed. If the user did NOT supply `platforms`, run ALL steps — every platform the skill supports.

STEPS:
  1. Tool `search_forums` — search_forums(query='{product} "how do i" OR "how to" OR "cant figure out"', time=year, get_sentiment=true)
     Collect recurring how-to questions; flag ones tagged confusion/fear as the most painful tasks. Set year (default year) and optional {region} to localize.
  2. Tool `search_reddit_comments` — search_reddit_comments(query='how do i {product}', sort_by=top, pages=3, get_sentiment=true)
     Add Reddit help-seeking comments and cluster everything by the underlying task. 3 controls how many pages to scan.
  3. Tool `search_twitter` — search_twitter(query='{product} "how do i" OR "how do you" OR "how to"', sort_by=relevance, pages=3, get_sentiment=true)
     Only run if {platforms} includes twitter: capture real-time help-seeking tweets and merge them into the task clusters; negative sentiment marks the most painful gaps.
  4. Tool `search_youtube` — search_youtube(query='{product} tutorial how to', pages=3)
     Count existing tutorials per clustered task; tasks with few or zero quality videos are content gaps.
  5. Tool `search_web` — search_web(query='<top_demand_task> {product} how to', time=year, include_ai_overview=true)
     Only run if {platforms} includes web: check whether a good doc, blog, or help-center page already ranks — and whether Google's AI overview already answers it. Strong existing coverage means lower deflection ROI; thin or missing coverage confirms the gap.
  6. Tool `search_youtube` — search_youtube(query='<top_demand_task> {product}')
     Confirm the single highest-demand question genuinely has no good video answer before prioritizing it; combine with the web coverage read to rank by deflection potential.

DELIVER: A prioritized list of doc/video topics where user demand is high (forums, Reddit, and X) but tutorial and search coverage is missing or thin (YouTube plus open web and Google's AI overview), ranked by deflection potential.

Note: each underlying tool call is billed at its normal endpoint price; get_sentiment adds a small per-page surcharge. Page through results as needed but stop once you have enough to deliver the outcome.