đ Works best with your applicant sheet. Connect Google Sheets and Jeremy AI reads each applicant for you.
Paste your must-haves and your notes on each candidate. Get one file back: a scorecard that checks every must-have, a ranking, one follow-up question per candidate and a claims check that shows where each fact came from. Nobody is contacted.
You'll see the SAMPLE must-haves and candidate notes pasted in, the scorecard written at 6x speed, the file saved with nobody contacted, the 4-column scorecard, the ranking of the candidates who meet every must-have, one follow-up question per candidate and the claims check that marks what is not provided.
What this use case does and who it's for
Hiring a setter is mostly sorting. You've got a handful of notes from calls and role-plays, and a gut feeling about each person. Gut feelings are where hiring goes wrong. You want every candidate held to the same bar, and you want nobody judged on anything but the job.
Jeremy AI reads your must-haves and your candidate notes and writes the scorecard for you, using only what's in your notes. The file has four parts:
A scorecard table. Four columns: candidate, role-play score, typing speed and "meets all must-haves" (yes or no).
A ranking. Only the candidates who meet every must-have, ordered by role-play score and then typing speed.
One follow-up interview question per candidate. Each is based only on your notes.
A claims check. Three columns: claim, source and status. Each fact is "stated in notes" or "not provided."
It's for agency owners, sales managers and coaches who hire setters or other entry-level sales roles. You stay in charge. It's a draft that helps you sort. You make the hiring decision. Jeremy AI doesn't message anyone or contact any candidate.
What you need before you start
A Jeremy AI account. Nothing to connect to try the sample.
Your must-haves, written as numbers where you can. The sample uses a role-play score of 3 or higher out of 5, a typing speed of 55 words per minute or higher, and a start date within 3 weeks.
Your notes on each candidate. The sample has four made-up candidates, shown by first name and last initial only: Mia T, Leo R, Ana S and Ben W. Each has experience, a role-play score, a typing speed and a start date.
A "not provided" note for anything missing. If you have no reference or quota numbers, say so. The sample claims check marks them "not provided."
Your rules for what the file must never do. The sample uses HARD RULES: every table has 4 columns or fewer, use only facts and numbers from the notes, never invent or round up a score, skill or experience, judge only on the must-haves and the notes (never on age, gender, background or personality), no claims about what most people experience, no discounts, bonuses, free trials or paying below the stated range, and never address you by name.
How to set it up
You use the default Jeremy AI worker. There's nothing to build.
Open Jeremy AI. From the dashboard, type in the main box. Check the worker selector says Jeremy AI.
Paste the prompt below. Swap in your own company, must-haves and candidate notes.
Click Send message.
Wait. The walkthrough runs this at 6x speed. In real time it took about 90 seconds.
When it says the file is saved, read the chat summary first.
Open the file card and read the scorecard.
Check each yes and no against your must-haves.
Read the ranking and the follow-up questions. Then read the claims check.
The prompt
This is the exact prompt from the walkthrough. The company, the candidates and every number are SAMPLE and made up. Replace them with yours.
Act as my "Hiring Scorecard" worker for SAMPLE Northstar Coaching, hiring one appointment setter (all data below is made up, candidates are shown by first name and last initial only). Build a scorecard from the SAMPLE candidate notes. Save ONE file named setter-scorecard.md with: 1) a SCORECARD table with exactly 4 columns (candidate, role-play score, typing speed, MEETS ALL MUST-HAVES yes or no), 2) a short RANKING of the candidates who meet all must-haves, ordered by role-play score then typing speed, 3) one follow-up interview question per candidate based only on the notes, 4) a CLAIMS CHECK table with exactly 3 columns (claim, source, status). HARD RULES: every table has 4 columns or fewer; use ONLY facts and numbers from the notes, exact numbers only, never invent or round up a score, skill or experience, and write "not provided" if something is missing; judge only on the must-haves and notes, never on age, gender, background or personality; no claims about what most people experience; never offer or mention discounts, bonuses, free trials or paying below the stated range; never address me by name anywhere in the chat reply or the file. Drafts only, nothing is sent and nobody is contacted.
MUST-HAVES: role-play score 3 or higher out of 5, typing speed 55 words per minute or higher, can start within 3 weeks.
CANDIDATE NOTES (SAMPLE):
1) Mia T: 14 months cold calling at a SAMPLE solar firm. Role-play 4/5. Typing 62 wpm. Can start in 2 weeks.
2) Leo R: no sales experience. Role-play 3/5. Typing 71 wpm. Can start immediately.
3) Ana S: 3 years inbound customer support. Role-play 5/5. Typing 48 wpm. Can start in 4 weeks.
4) Ben W: 8 months as a setter at a SAMPLE agency. Role-play 2/5. Typing 55 wpm. Can start in 1 week.
Prompting tips and how to adapt it
Write must-haves as numbers. "Role-play 3 or higher out of 5" can be checked. "Good attitude" can't.
Keep notes short and factual. One line per candidate, with the same items in the same order.
Use first name and last initial. It keeps the sorting about the notes.
Keep the fairness rule. The prompt says to judge only on the must-haves and the notes, never on age, gender, background or personality. Don't take it out.
Don't let it guess. Say "never invent or round up a score, skill or experience." Then a missing item says "not provided."
Ask for the exact order. The sample ranks by role-play score, then typing speed. Tell it your tie-break rule.
Ask for one follow-up question each. A single question based on the notes is easy to use in an interview.
Keep the tables to 4 columns or fewer. That's why the scorecard fits the file panel with nothing cut off.
Ask for the claims check. It shows which facts came from your notes and which are not provided.
Say drafts only. The prompt says nobody is contacted, so nothing goes out.
Adapt it. Use it for closers, support reps or assistants. Swap in your own must-haves.
The result to expect
One Markdown file named setter-scorecard.md, titled "Appointment Setter Hiring Scorecard (SAMPLE)." It opens with a note that it's sample data, drafts only and judged only on the stated must-haves and the candidate notes. With the sample notes you should see:
A must-haves line. Role-play score 3 or higher out of 5, typing speed 55 wpm or higher, can start within 3 weeks.
A 4-column scorecard.
Mia T: role-play 4/5, 62 wpm, meets all must-haves: Yes.
Leo R: role-play 3/5, 71 wpm, meets all must-haves: Yes.
Ana S: role-play 5/5, 48 wpm, meets all must-haves: No.
Ben W: role-play 2/5, 55 wpm, meets all must-haves: No.
Notes on misses, from the notes only. Ana S misses on typing speed (48 wpm, below 55) and start time (4 weeks, beyond 3 weeks). Ben W misses on role-play score (2/5, below 3).
A ranking of the two who meet every must-have. Mia T is first, with role-play 4/5, typing 62 wpm and a start in 2 weeks. Leo R is second, with role-play 3/5, typing 71 wpm and an immediate start. The file notes that Leo R has no sales experience listed in the notes.
One follow-up question per candidate. For example, Mia T's asks how she'd adapt her cold call opening for a completely different industry on day one. Leo R's asks what he did to prepare for the role-play and how he'd close the experience gap in his first 30 days. Each question uses only what's in the notes.
A 3-column claims check. Each candidate's experience, score, typing speed and start date is "Stated in notes," with the source "SAMPLE candidate notes." References, certifications or prior quota numbers for any candidate are "Not provided."
The chat reply summarizes the file and offers quick follow-ups, like adding a second question per candidate. Nothing is sent and nobody is contacted.
These are sample results from made-up candidates. Wording changes from run to run. Read every line before you use it. Check each yes and no against your must-haves. Check that every follow-up question is about the job and the notes. The scorecard is a sorting tool. You make the hiring decision.
Common mistakes and fixes
A yes or no doesn't match your must-haves. Say: "Recheck every candidate against each must-have and show the numbers."
It comments on age, gender, background or personality. Delete it. Remind it: "Judge only on the must-haves and the notes."
A score or skill is invented or rounded up. Delete it. Remind it: "Use only the notes. Never invent or round up a score, skill or experience."
Something missing is filled in. Say: "Write 'not provided' for anything the notes don't say."
A follow-up question isn't based on the notes. Say: "Base each question only on that candidate's notes."
The ranking includes someone who missed a must-have. Say: "Rank only candidates who meet all must-haves."
The tie-break isn't what you want. Tell it: "Order by role-play score, then typing speed," or your own rule.
A pay offer or perk appears. Delete it. Remind it: "Never offer or mention discounts, bonuses, free trials or paying below the stated range."
The table has more than 4 columns. It may clip in the file panel. Remind it: "Every table has 4 columns or fewer."
Real candidates in a test. For practice, use made-up candidates and label them SAMPLE. When you use real candidates, keep their details private.
You treat the scorecard as the decision. Don't. It's a draft to help you sort. You decide.