Prompt Chain: Screen Notes to Submitted Candidate in 4 Automated Steps

What This Builds

A four-step prompt chain that transforms raw, messy screen notes into a complete submission package — submittal letter, candidate summary for internal use, interview coaching document for the candidate, and a client-ready one-liner for email subject lines — in under 5 minutes. Each prompt takes the output of the previous one as its input, so you're not re-explaining the candidate's background four times. You run through the chain once per candidate, end up with everything you need, and send.

Without this chain: you write a submittal (20 minutes), manually write a separate internal summary if a colleague asks (10 more minutes), improvise a coaching document for the candidate (another 15 minutes), and write the email subject line from scratch. Total: 45–50 minutes per placement cycle.

With this chain: 4 prompts, 5 minutes, everything done.

Prerequisites

• Claude Pro ($20/month) with a configured Recruiting Project (see Level 4 guide: "Custom Claude Project") — OR ChatGPT Plus ($20/month) with custom instructions configured
• A consistent screen note format (even rough bullets work — see the template in Part 1)
• The job description or at minimum a bullet list of the client's key requirements

The Concept

A prompt chain is just a series of prompts where each one builds on what came before. It's the difference between asking a helper "write me a summary" cold versus handing them a document and saying "now, based on what you just read, write me an email about it." The second approach produces better output because the helper already has the context.

The trick with prompt chaining is that you feed the AI output from Step 1 into Step 2 as input — so you're not re-entering data. You copy the submittal output from Step 1 and paste it at the top of Step 2's prompt. The AI now has the full candidate story already built when it writes the coaching document. By Step 4, it's essentially working from a complete picture of the candidate it built itself.

Build It Step by Step

Part 1: Standardize Your Screen Note Format (15 minutes)

The chain works best when your screen notes follow a consistent structure. You don't need complete sentences — bullets are fine. Here's a template you can use or adapt:

SCREEN NOTES TEMPLATE

Candidate name (first name + last initial is fine):
Current title / company:
Years of relevant experience:
Key skills (be specific — software, certifications, equipment):
Team/direct report size (if supervisory role):
Notable achievement or differentiator (what makes them interesting):
Availability (start date):
Compensation ask:
Reason for looking:
Interest level (1–5):
Location / commute (any issues?):
Red flags (if any):
My recommendation: SUBMIT / HOLD / PASS

Spend 5 minutes after each screen filling this in. Even rough notes ("7 yrs WH exp, SAP, managed 20, available 2 wks, $70K, laid off, solid, submit") are enough for the chain to work.

Part 2: Prompt 1 — Generate the Candidate Submittal

Open your Claude Project or a ChatGPT session with your custom instructions active. Paste this prompt, filling in the bracketed sections:

PROMPT 1: CANDIDATE SUBMITTAL

Here are my screen notes for a candidate I want to submit:

[PASTE YOUR SCREEN NOTES HERE]

Job requirements (key criteria the client cares about most):
[PASTE THE JD BULLETS OR TYPE 4–6 KEY REQUIREMENTS]

Write a 3-paragraph candidate submittal to send to the client.

Paragraph 1: Lead with the candidate's single strongest qualification and connect it explicitly to the client's stated priority.
Paragraph 2: Cover secondary qualifications, relevant skills, and any differentiating factor.
Paragraph 3: Availability, compensation, interest level, and recruiter recommendation. Be direct.

Keep total length under 200 words. Professional tone. No hollow phrases ("strong communicator," "team player") without specific evidence. Use the candidate's years of experience and specific skills — not general descriptions.

Label this output: [SUBMITTAL]

Review the output. If it's off, give it one correction ("The second paragraph is too long — cut it to 2 sentences and focus on the SAP experience"). Once you're satisfied, copy the entire output for use in Step 2.

Part 3: Prompt 2 — Generate the Internal Candidate Brief

Paste this into the same chat conversation (so Claude still has the full context), adding the submittal you just generated:

PROMPT 2: INTERNAL CANDIDATE BRIEF

Here is the submittal I just generated for this candidate:

[PASTE THE SUBMITTAL OUTPUT FROM PROMPT 1]

Now write a short internal brief I can send to a colleague or another recruiter on my team. This is not client-facing — it's internal.

Format:
- Name: [First name + last initial only]
- Role background: [1 sentence]
- Key skills: [3 bullets, brief]
- Availability: [brief]
- Compensation: [brief]
- Fit note: [1 sentence — why this candidate is interesting or what to watch for]

Keep it under 100 words. Conversational, not formal.

Label this output: [INTERNAL BRIEF]

Part 4: Prompt 3 — Generate the Candidate Interview Coaching Document

Still in the same conversation, paste:

PROMPT 3: INTERVIEW COACHING DOCUMENT

Using what you now know about this candidate and the job requirements, create an interview coaching document I can send to the candidate before their interview.

Include:
1. The top 5 questions the hiring manager is likely to ask for this role, based on the requirements
2. For each question, a 2-sentence coaching note on what a strong answer looks like for this candidate specifically — reference their actual experience where possible
3. One behavioral question in STAR format with a coaching note on how this candidate should frame their answer
4. A 3-bullet "what they care about" summary of what the client is really evaluating

Format it cleanly. The candidate will receive this directly, so use "you" — address the candidate, not me.

Do not include the candidate's compensation or availability information in this document.

Label this output: [COACHING DOCUMENT]

Part 5: Prompt 4 — Generate the Submission Email Subject Line and Opening Line

Final step. Still in the same conversation:

PROMPT 4: EMAIL SUBJECT AND OPENING LINE

Using everything above, write:

1. Three options for the email subject line I'll use to send the submittal to the client. Each subject line should be under 60 characters, include the candidate's title/background, and create a reason to open the email. Do not use generic lines like "Candidate for Your Review."

2. One opening sentence for the email body that leads directly into the submittal. One sentence. Make it specific to this candidate — not a generic introduction.

Label these outputs: [EMAIL SUBJECT OPTIONS] and [EMAIL OPENER]

Real Example: Marcus G., CNC Machinist Placement

Screen notes (what the recruiter typed after the call):

Marcus G — CNC setup/run, 9 yrs total, 5 yrs Haas (VF series), Fanuc exposure, reads blueprints independently, GD&T literate, currently at a tier-2 auto supplier doing mixed model production — been there 4 years, layoff announced last month. Looking NOW. $28/hr ask. Will take $26 to start somewhere stable. 2nd shift preferred (wife works days), available immediately. Solid guy, asked good questions. SUBMIT.

Job requirements: Haas setup machinists, blueprint reading required, $26–$30/hour, 2nd shift, precision metal fab shop, immediate need.

Prompt 1 output (SUBMITTAL):

Prompt 2 output (INTERNAL BRIEF):

Prompt 3 output excerpt (COACHING DOCUMENT, first 2 questions):

Prompt 4 output (EMAIL SUBJECT OPTIONS):

Email opener:

Total time: Screen call done, notes typed, 4 prompts run in sequence — complete submission package ready in under 8 minutes.

What to Do When It Breaks

• Prompt 2 output ignores the submittal you pasted → You may have accidentally started a new chat session. Check that you're still in the same conversation thread. If the AI says "I don't see a submittal" — paste it again.
• Coaching document is too generic ("answer behavioral questions with STAR format") → The AI didn't have enough candidate-specific information to personalize the coaching. Go back to your screen notes and add one or two specific examples the candidate mentioned on the call. Even brief ("mentioned he caught a dimensional issue on a critical bore last year — saved the batch") gives the AI something to work with.
• Submittal is too long → Add the word limit explicitly ("Keep it under 200 words") in Prompt 1. If it's still long, try adding "Use short, declarative sentences. No compound sentences."
• Output sounds robotic → This usually means your screen notes are too sparse or structured. Add a few conversational words in your notes about what stood out ("solid guy, asked about growth path — seems genuinely interested") and the AI will pick up on that texture.
• You're in a basic Claude/ChatGPT free account without Projects → The chain still works, it's just slower because you have to paste context in at the start. Add this block to the top of Prompt 1: "I am a staffing recruiter at a light industrial agency. My clients are manufacturing and distribution companies. I write 3-paragraph candidate submittals in professional but direct language." Then run the chain as normal.

Variations

• Simpler version: Just run Prompts 1 and 3 — get the submittal and the coaching document. Skip the internal brief and subject line optimization until you're comfortable with the chain.
• Extended version: Add a Prompt 5 after the chain: "Based on this candidate's background and the job requirements, write 3 counter-offer objection responses I can prepare in case the candidate receives a retention counter-offer from their current employer." Now your placement preparation is complete end-to-end.

What to Do Next

• This week: Run your next 3 submittals through the full chain. Time yourself. Compare the output quality to your manual drafts.
• This month: Add the chain to your Claude Project (see Level 4 guide: "Custom Claude Project") so it starts from the context of your agency, clients, and preferred formats — outputs will be noticeably more calibrated.
• Advanced: Add a fifth prompt that writes the candidate's formal offer summary email once a placement is made, using the compensation and start date from your screen notes. Now the chain covers entry-to-offer.

Advanced guide for staffing recruiter professionals. These techniques use more sophisticated AI features that may require paid subscriptions.