Applicant review
Paused while US hiring is on hold until mid-November 2026.
A weekday run that logs everyone who applies for our sales roles, scores them on one rubric, and drafts the replies and call times.
- Why
- One job post brought 66 applicants. Reading profiles one by one was eating my mornings, and I wanted every candidate judged on the same bar.
- How long
- Built 6 Oct 2026; first run 12 Oct
How I built it
- I wrote a 100-point rubric with two must-have gates, and the cutoffs: 70 and up gets a call, 55 to 69 a written screen, under 55 a polite no.
- Claude in Chrome reads the LinkedIn applicant list, read-only. It never clicks Reject or Message, because those can message the person.
- Applications that arrive by email are added too, and everything is de-duplicated into one Google Sheet.
- Replies become Gmail drafts copied to my business partner; LinkedIn notes go into the messages sheet for me to send.
- Names stay in the sheet only. My notes keep counts.
- To copy it you need
- Claude, or the AI of your choice, with Claude in Chrome for LinkedIn, a Gmail connector and a Google Sheet. The gotcha: score before you offer times. Once an email has gone out, stand behind it.
Try it yourself
Copy this prompt and paste it into ChatGPT or Claude to try it yourself.
I'm looking at Shweta Kamble's applicant review on whatarewebuilding.ai. Every weekday it reads the LinkedIn applicant list read-only, logs each person in one Google Sheet, scores them on a 100-point rubric with two must-have gates (70 and up gets a call, 55 to 69 a written screen, under 55 a polite no), and drafts the replies. Help me build a starter version for a role I'm hiring. Ask me what the job really needs, then write the rubric and the run's instructions. The gotcha Shweta hit: LinkedIn rated almost everyone a 'Top fit', so ignore labels and read the work history. Score before offering any times. Ask me for anything you need.
Tools I used for this project












