Overview

Lecture URL: https://skill.re/learn/recruiting/iterating-with-ai-following-up-clarifying-and-refining.php

TRANSCRIPT: Iterating with AI: Following Up, Clarifying, and Refining

Course: AI for Recruiters - Professional Credential

Module: Level 2: Hands-On Foundations

Section: Chapter 6 -- Effective Prompting for Recruiting Tasks

Theme: effective-prompting-for-recruiting-tasks

Lecture: 6.3

Duration: 60 min

Format: Workshop + Hands-On

Audience: Recruiters beginning to use AI tools

Prerequisites: L1 Certification

What you will learn: You'll master the prompting loop: how to review AI outputs, identify gaps,

ask follow-up questions, and iterate toward increasingly useful results. By the end, you'll be able

to refine outputs through structured feedback and get AI to provide exactly what you need.

INTRODUCTION

You send a prompt to AI. You get back an output. You read it and think: "This is close, but it's

missing something." Or: "This is good, but can you go deeper here?" Or: "This is useful for

screening, but I also need to know..."

That's where iteration comes in. The best recruiting prompts are rarely one-shot. They're

conversations. You ask, you get back an output, you give feedback, and the next version is better.

Today, we're learning how to be an effective conversationalist with AI.

THE ITERATION LOOP

Iteration has four steps: review, identify gaps, ask follow-ups, and refine. Let's walk through

each.

STEP 1: REVIEW THE OUTPUT

Read what AI gave you, but read it like a critic. Ask:

• Is this accurate? Does it align with what I know about the candidate or situation?
  - Is it complete? Did it answer all parts of my question?
  - Is it useful? Can I actually use this for my decision-making?
  - Is it biased? Does it make unfair assumptions or encode discrimination?
  - Is it the right format? Can I scan it quickly, or is it hard to find what I need?

When you review, you're not looking for perfect. You're looking for: "What would make this more

useful?"

STEP 2: IDENTIFY THE GAPS

Be specific about what's missing. Don't just think "this isn't quite right." Instead:

• "The output covers years of experience but doesn't address technical depth in specific areas I care about."
  - "This summary tells me what they did, but not why they did it or what they learned."
  - "This compares candidates on experience but doesn't address soft skills or communication ability."
  - "This identifies red flags but doesn't suggest questions to ask the candidate."

Specific gap identification makes your follow-up prompt much stronger.

STEP 3: ASK FOLLOW-UP QUESTIONS

A follow-up prompt is short and targeted. You're not re-doing the entire analysis. You're asking

AI to go deeper on a specific dimension.

Example: You asked AI to summarize a candidate's background. It gave you a good summary, but it

didn't address their leadership experience. Your follow-up:

"Thank you. Now focus specifically on: (1) Teams they've led or managed, (2) Evidence of

mentoring or developing others, (3) How they describe their leadership style or approach. If this

information isn't clearly in the resume, note that."

That's a targeted follow-up. It tells AI exactly what to focus on.

Another example: You asked AI to identify red flags in a resume. It flagged "job hopping" as a

concern. You want to understand this better. Your follow-up:

"I see you flagged frequent job changes as a red flag. For each role, what's the tenure? And

what might explain the changes (company acquisition, role expansion, geographic move, or something

else)? Is there a pattern, or are these one-time occurrences?"

This transforms a flag into understanding. You're not dismissing the candidate -- you're getting

context.

STEP 4: REFINE THE PROMPT FOR NEXT TIME

After you iterate once or twice, you've learned something. Update your prompt template so the next

time, you get better output immediately.

Example: You've been screening engineers, and you keep having to follow up on "What specific

projects have they shipped?" Add that to your prompt template:

"For each role, identify: (1) Specific projects they shipped or contributed to, (2) Technologies

they used, (3) Team size, (4) Time-to-market or impact metrics if available."

Now every engineer screening includes project details without needing follow-up.

THE ANATOMY OF STRONG FOLLOW-UP QUESTIONS

Follow-up questions should be brief, focused, and build on what AI already knows. Here's the

anatomy:

1. Acknowledge what AI got right: "Thank you for the summary on technical experience."
2. Identify the gap: "I want to dig deeper on leadership and team impact."
3. Ask specifically: "Focus on: (1) Teams led or influenced, (2) Feedback from direct reports if

mentioned, (3) Any evidence of developing junior team members."

1. Add constraints if needed: "If this information isn't explicit in the resume, note that rather

than inferring."

That structure keeps your follow-up focused and productive.

ITERATION IN DIFFERENT RECRUITING CONTEXTS

CONTEXT 1: SCREENING AND ANALYSIS

You're asking AI to analyze a resume or interview. Iteration looks like:

First prompt: "Summarize this resume."

Follow-up 1: "Now focus on technical depth. What technologies does this person know deeply?"

Follow-up 2: "Does anything suggest they've worked in cross-functional teams? Give specific examples."

Follow-up 3: "Red flags to explore in an interview?"

By the third iteration, you have a complete picture.

CONTEXT 2: DRAFTING AND REVISION

You're asking AI to draft an outreach email or job posting. Iteration looks like:

First prompt: "Draft an outreach email to this candidate."

Feedback: "This feels a bit generic. Can you personalize it more to their specific experience?"

Feedback 2: "Better. But I want the tone to be slightly warmer. Can you make it feel like a

conversation between peers?"

CONTEXT 3: COMPARISON AND CALIBRATION

You're asking AI to help you compare candidates. Iteration looks like:

First prompt: "Compare these three resumes."

Feedback: "Good comparison. Now tell me: which person would handle ambiguity best? What evidence

supports that?"

Feedback 2: "Helpful. One more thing: which person would need the most onboarding, and which could

hit the ground running?"

Notice in all three contexts: you're not rejecting or dismissing the output. You're deepening it.

WHEN TO ITERATE VS. WHEN TO MOVE ON

Iteration is powerful, but it has diminishing returns. Know when to stop.

You should iterate if:

• The output is mostly useful but missing a dimension you care about.
  - The output is close but needs adjustment in tone, format, or emphasis.
  - You want to compare approaches or get a second perspective on the same information.

You should move on if:

• The output is useful enough for your current decision.
  - More iteration won't change your decision about the candidate.
  - You're spending more time iterating than the output is worth.

Smart iteration is about leverage. If one follow-up question will save you 30 minutes of analysis,

do it. If you've already spent 15 minutes iterating on something that's 80% good, move on.

ANTI-PATTERNS

ANTI-PATTERN 1: VAGUE FEEDBACK

Description: Telling AI the output is "not quite right" without saying what's missing.

Example: "This summary feels off. Can you try again?"

Why it fails: AI has no idea what to improve. You'll probably get a similar output with different

wording. You're wasting both your time.

How to avoid: Always be specific about gaps. "This tells me about technical skills but not about

leadership experience" is much better than "This feels off."

ANTI-PATTERN 2: OVER-ITERATING

Description: Asking for revision after revision after revision.

Example: You've asked for three follow-ups, each one improves the output 5%, and you keep asking

for more.

Why it fails: You hit diminishing returns fast. After 2-3 iterations, you're usually getting 95%

of the value. More iteration is wasted effort.

How to avoid: After 2-3 iterations, step back. Is the output actionable? If yes, move on. If no,

you might need a different prompt approach entirely, not more iteration.

ANTI-PATTERN 3: LETTING AI DRIVE THE ITERATION

Description: Asking AI to suggest follow-ups or "What else should I consider?"

Example: "You've summarized the resume. What other questions should I ask about this candidate?"

Why it fails: You're letting the AI agenda drive your recruiting decision. You should know what

information matters for your hire -- don't delegate that to AI.

How to avoid: Know what you're evaluating for. You identify gaps. You ask follow-ups. AI supports

your decision-making, not the other way around.

PRACTICE PROMPTS

Exercise 1: Identify Gaps in an Output

Take a summarized resume or interview note from AI. Review it and list: (1) What's included and

accurate? (2) What's missing that you'd want to know? (3) What format changes would make it more

useful? (4) What bias or assumptions does it make?

Exercise 2: Write a Follow-Up Question

Based on the gaps you identified, write a follow-up prompt that addresses 2-3 of them. Make it

specific and focused.

Exercise 3: Iteration Test Run

Choose a recruiting task (screening, sourcing, drafting). Write an initial prompt, get an output,

and do 2-3 follow-up iterations. Document: What did each iteration add? At what point did you have

enough information?

Exercise 4: Know When to Stop

You've asked AI to draft an outreach email. First version: decent, but generic. Second version:

better, more personalized. Third version: slightly different wording but same content. At what

point would you stop iterating? Why?

Exercise 5: Refine Your Template

Take a recruiting task you do regularly. Write your initial prompt. Test it. Iterate twice. Then

write the refined prompt that captures what you learned. This becomes your new template.

KEY TAKEAWAYS

1. Iteration is a conversation with AI. You ask, you review the output, you identify gaps, you ask

follow-ups, and you refine for next time. Most strong recruiting outputs come from 2-3 iterations,

not one-shot prompts.

1. When you review output, be a critic. Ask: Is it accurate? Complete? Useful? Unbiased? In the

right format? Specific gap identification makes your follow-ups much stronger.

1. Follow-up questions should be brief, focused, and build on context. Acknowledge what's good,

identify the specific gap, ask for the missing piece, and set constraints if needed.

1. Iteration has diminishing returns. Usually 2-3 follow-ups get you 90%+ of the value. After that,

you're wasting time. Know when to move on.

1. You drive the iteration. You identify what matters to your hiring decision. You decide what

information is missing. AI supports your decision-making, not the other way around.

1. Refine your prompts based on iteration. When you find yourself asking the same follow-up

question over and over, build it into your template.

GLOSSARY

Iteration Loop: The cycle of asking a question, reviewing the output, identifying gaps, asking

follow-ups, and refining the prompt. Most effective recruiting prompts develop through iteration.

Follow-Up Question: A targeted, brief prompt that deepens one dimension of a previous output. Good

follow-ups are specific and build on context already established.

Gap: Information or analysis that's missing from an AI output. Identifying gaps clearly and

specifically makes follow-up questions much more effective.

Diminishing Returns: The point at which additional iterations produce minimal improvement in the

output. Smart iteration stops before diminishing returns become too steep.

Prompt Refinement: Updating your prompt template based on what you learned during iteration. This

ensures future outputs are better without needing as many follow-ups.

SYNTHESIS AND APPLICATION

Effective recruiting with AI is iterative, not one-shot. You don't need perfect prompts -- you need

prompts that are good enough to start with and that you can refine through conversation. This is

actually more human and more nuanced than trying to write the perfect prompt the first time.

This week, pick one recruiting task you do regularly. Write a basic prompt. Use it twice, and

iterate each time. Notice how the output gets better. Then write the refined version that

incorporates what you learned. That becomes your template for the next 10 times you do this task.

REFLECTION EXERCISE

1. Think about a recruiting task you've done recently. If you had iterated with AI instead of

accepting the first output, what would you have wanted to explore further?

1. What recruiting decisions require the deepest information? Which decisions could you make with

less? How might iteration depth depend on decision importance?

1. How do you currently know when you've gathered "enough" information about a candidate? How might

AI help you structure that?

1. When you review AI outputs, do you find yourself assuming it's "good enough" too quickly? What

would change if you treated every output as a draft that could be improved?

CLOSING REMARKS

Iteration transforms AI from a tool that gives you answers into a tool you have a conversation with.

That conversation is where the real insight happens. In our next section, we'll apply these prompting

skills to one of the highest-stakes recruiting tasks: drafting authentic, personalized outreach.

AI for Recruiters Certification Program

Level 2: Hands-On Foundations | Effective Prompting for Recruiting Tasks | Lecture 6.3

A SkillsClinic initiative.

Duration: ~60 minutes | Word Count: ~2,380