Qualify LinkedIn connections for outreach with AI

Overview

Qualify LinkedIn connections for outreach with AI when you already have a large network but no fast way to decide who is worth contacting. This playbook keeps the qualification step inside a Cockpit spreadsheet, so the imported LinkedIn fields, the AI decision, and the next outreach action all stay visible on the same rows instead of disappearing into separate tools. The result is a simple approval layer for warm network prospecting, where you can see exactly which connections passed the fit check before you write a message or start a sequence.

Import your LinkedIn connections export, score each row with an AI true-or-false fit check, and turn your existing network into a cleaner outreach segment that is easier to review and act on.

Step-by-step process

1. 1

   Export the LinkedIn connections file you actually need

   The goal of this step is to get your existing network out of LinkedIn and into a format that Cockpit can score row by row. In the video, the operator goes to LinkedIn settings, opens the Data Privacy area, requests the download archive, then waits for the email that contains multiple CSV files.

   What matters here is selecting the connections CSV, not importing the full archive blindly. That file gives you the starting rows for the workflow. If the archive contains extra CSVs that are not useful for prospecting, leave them out so the spreadsheet stays focused on the contacts you may want to review.

2. 2

   Import the CSV into a new Cockpit spreadsheet

   Create a new spreadsheet and choose a CSV source. The walkthrough shows the operator previewing the archive contents, picking the connections file, and bringing the available fields into Cockpit so they become visible columns in the sheet.

   This step matters because the later AI decision is only as useful as the row context you preserve. Good output at this stage is a clean spreadsheet where each row represents one connection and the person or company fields you care about are present before you add any AI columns.

3. 3

   Create a boolean AI qualification column

   Add an AI answer column and phrase the prompt as a narrow qualification decision. In the video, the operator asks whether the person and company are a good target for SEO services, says the target roles are heads of marketing, GTM people, and founders, and then tells Cockpit to return only true or false.

   The important detail is the strict output format. A boolean result is easier to filter than a paragraph. If the model starts returning soft explanations or ambiguous phrasing, tighten the prompt before you run the workflow at scale.

4. 4

   Test the logic on a small batch before running the full sheet

   The demo runs the first ten rows rather than processing the whole file immediately. That is a practical quality check: you can see whether some rows are marked true and others false, verify that the model is reading the source fields sensibly, and catch prompt drift early.

   If the results look too strict or too generous, revise the fit criteria before continuing. This is the easiest point to fix bad targeting, because you still have a small sample in front of you instead of hundreds of rows to clean up later.

5. 5

   Filter the approved rows and move into follow-up

   Once the true-or-false column is populated, use the true rows as your approved outreach segment. The video ends by pointing to the next two actions: generate a first cold message for the relevant people or send the approved rows into Cockpit's sequence builder.

   That makes this workflow especially useful when you want a reviewable bridge between sourcing and action. The sheet keeps the original LinkedIn data, the AI decision, and the next campaign step together, so you can filter, review, and reuse the workflow later.

Key outputs