Find First-Hand Experience

Finding real people talking from real experience used to mean scrolling endlessly through search results, forums, and social threads—an exercise that’s getting tougher as AI-generated and affiliate-stuffed pages crowd the web. Google’s push toward E-E-A-T shines a spotlight on “Experience,” but it doesn’t hand you an easy way to surface it. That’s what the “Find First-Hand Experience” app tackles: it hunts down pages where someone actually tried, tested, or lived the thing you’re researching, so you can cut through the noise and trust what you read.

Walkthrough: How the App Works

Input

SME Topic Query – one line describing what you want to learn (e.g., “cold plunge benefits,” “Switch OLED screen issues”).

Step 1 – Build a laser-focused query

A GPT-4o Chat Model reviews your prompt and crafts an advanced Google query in under 32 words. The prompt bakes in smart operators like ("author" OR "written by"), personal-pronoun filters ("I" OR "my"), and negative keywords (-affiliate -commission) to sidestep corporate content. It also drops a single -site: exclusion to avoid the brand likely dominating the SERP. The result is an operator-rich string ready for copy-pasting straight into Google—no fluff, no extra characters.

Step 2 – Pull the SERP

The SERPs Search node sends that query to Google and retrieves up to 50 web results with titles, snippets, and links. This gives us raw material that (a) already mentions real people and (b) isn’t polluted by the site we excluded.

Step 3 – Surface the humans first

Another Chat Model sorts the JSON. Listings that show a named author in the rich snippet float to the top, then those with a name in the regular snippet, followed by everything else. Columns like thumbnail and domain are trimmed to keep things clean, and duplicates are removed so each source appears once.

Step 4 – Curate the winners

The final Chat Model turns the sorted JSON into a concise Markdown list of the ten most helpful results. For each, it extracts:

• Title and author (bold-numbered)

• Full URL

• A one-line summary explaining why this is genuine first-hand experience (“John Doe tried the supplement for 30 days and reported improved sleep…”)

The list is automatically ranked for relevance to your original query, and irrelevant or redundant pages are quietly dropped. What you see in the Text output is ready to paste into a brief, Slack message, or content outline.

Ways to Customize

• Tighten or loosen author signals – Edit Step 1’s prompt to include or remove terms like “guest post,” “case study,” or niche forum names.

• Change sentiment focus – Add “bad experience” or “pros and cons” to the prompt if you want critical takes instead of glowing reviews.

• Expand sources – Swap -affiliate for -amazon or -reddit depending on where spam creeps in for your industry.

• Use News or Scholar search – In Step 2, flip search_type to "news" or "scholar" to surface journalistic investigations or academic studies with named authors.

• Output formats – Switch Step 4 to JSON if you plan to feed the curated list into another workflow or dashboard.

Running at Scale with Bulk Runs

Bulk processing like this is a great example of content automation in action, streamlining your workflow while handling large volumes of data.

1. Export a CSV with a column titled SME Topic Query containing all the subjects you need first-hand insights for.

2. Go to Bulk Runs in Moonlit, select the “Find First-Hand Experience” app, and upload your CSV file.

3. Map the CSV column to the app’s input field, hit Run, and grab your compiled Markdown lists once processing finishes.