LinkedIn creator sourcing and vetting workbench for B2B

I run B2B influencer campaigns and I want a sourcing and vetting workbench built on the one dataset almost nobody uses: the LinkedIn posts creators have publicly disclosed as paid brand partnerships. Build me an app for finding, judging, and shortlisting creators before I ever contact them.

The main view is a search screen backed by the LinkedIn Ad Library Search Brand Partnerships operation. I search two ways: by topic keyword, so I can ask who has taken money to talk about data infrastructure this year, and by brand name, so I can ask which creators my competitors have already paid. The operation returns disclosed posts, so the handler should group them into one row per creator. Each result row shows the creator name, how many disclosed sponsored posts matched my search, the distinct brands they have posted for, and the date of their most recent disclosed post. Sort by recency by default and let me re-sort by volume. Always show the underlying matching post count so I can see whether the evidence is thick or thin.

Opening a creator row shows that creator's full disclosed partnership history, fetched with the same Search Brand Partnerships operation scoped to them. List every disclosed post with its date, the brand partner, and a snippet of the post text. Above the list, show the summary numbers the decision actually rests on: total disclosed sponsored posts, how many distinct brands they have served, roughly how many disclosed sponsored posts per month over the covered window, how recently they last posted sponsored content, and a read on how heavily their sponsored copy skews promotional.

Compute two flags on every creator and render them prominently on both the result rows and the detail view, because these decide most of my calls. The conflict flag fires when a creator has a disclosed post for any brand on my competitor list, and it should name which competitor and link to the specific post. The saturation flag fires when sponsored posts make up more than my configured share of their recent disclosed activity, because an over-sponsored creator converts badly. The competitor list and the saturation threshold live on a settings screen I can edit, and changing either recomputes the flags on what is already on screen.

Shortlists persist in General Input Database. I want named shortlists, one per campaign, and I add creators to them in one click from either the search results or the creator detail. Each entry stores the creator identity, the shortlist name, my own free-text notes, and a status of considering, reached out, booked, or passed that I can change inline. Give me a shortlist view listing my campaigns, and inside one campaign, the creators with their flags, status, notes, and vetting verdict. Create the tables on first load if they do not already exist.

Every creator has a "Vet this creator" button that kicks off a background agent inside the app. The agent reads that creator's disclosed posts and produces a written verdict in four parts: how well their disclosed subject matter fits my category, a summary of the brands they already serve and what that says about their positioning, an explicit call-out of any conflict with my competitor list, and a drafted outreach angle that references specific posts of theirs rather than generic flattery. The agent writes the verdict back onto that creator's row in General Input Database so the app displays it inline, with a timestamp and a visible running state while it works.

When a shortlist is ready I hit Export and the app builds a shareable document using the Google Docs Create Document operation for the file and Batch Update Document to fill it in. Title the doc with the campaign name. For each creator include their name, their flags, the brands they have disclosed partnerships with, my notes and status, and the full vetting verdict. Show me the document link when it finishes and store it on the shortlist so I can find it again later.

One rule has to show in the interface and not just in the code: this corpus only contains posts a creator explicitly tagged as a paid partnership. It is a disclosure record, not a complete content history, and the search screen should say so plainly instead of implying full coverage. Follower counts and engagement metrics are not available from this source at all, so never invent them, never render empty audience-size fields, and make the flags and the vetting lean entirely on disclosed partnership patterns. Where the evidence is thin, say the evidence is thin.

The user is an influencer marketing lead running several campaigns at once, so open the app on whichever shortlist I touched last. Handle empty search results gracefully with a note that no creator has disclosed a partnership matching that term, which is common for narrow keywords, and suggest broadening the keyword or searching by a competitor's brand name instead.