AI Academy/Prompt library/Reporting and process
Reporting and process prompts for ChatGPT, Claude, Grok and Gemini
Six prompts for the weekly hiring manager update, finding where the process stalls, AI ground rules, desk metrics, profitable role types and market maps.
- ChatGPT
- Claude
- Grok
- Gemini
- 6prompts
- 4AI assistants
- Freeplans work fine
Updated September 2026
Before you paste
- Paste an export or your own figures; the AI should not supply numbers.
- Ask it to show the working behind every figure.
- Use only an account your company approves for ATS exports.
In short
Reporting is where AI saves the most time for the least risk, as long as the numbers come from you. Paste the export or the figures, ask for the working, and let the assistant turn them into an update a hiring manager reads in a minute.
Keep candidate data outTake names and contact details out before you paste anything into a personal account. Module 4, Lesson 5 explains why.
Read before it goes outCheck every answer before it reaches a candidate or client. The AI drafts; you decide who is hired and who is not. Keep the last line of each prompt: it makes the assistant flag missing details instead of inventing them.
Reporting and process prompts
6 promptsThese pair with the Module 8 lesson on analyzing recruiting data. Ask for the working behind every number so you can check it, and paste exports only into an account your company approves.
Write the weekly hiring manager update
Use whenIt is Friday and several managers are waiting to hear where their roles stand.
Here is this week's pipeline for [job title]: [paste the number at each stage, what moved, interviews booked, blockers]. Write an update for the hiring manager that starts with the one thing they most need to know, covers what moved this week and what happens next, asks clearly for anything I need from them with a date, and names each risk with the action I propose. Under 200 words, no recruiting jargon, no candidate surnames. Use only the details I give you here. If something you need is missing, write [missing] and tell me what to add, rather than making it up.
Watch outCheck every number against your ATS before it goes. A wrong count costs more trust than a late update.
Find where your hiring process stalls
Use whenTime to hire keeps creeping up and nobody can say why.
Here are my hiring numbers for [period]: [paste the count at each stage and the average days spent in each]. Work out: 1) the conversion rate between each pair of stages, showing the sums; 2) where the most people drop out, with two likely reasons for each; 3) which stage adds the most days while changing the fewest decisions; 4) three fixes ranked by effort against likely impact; 5) the number I should watch to know whether a fix worked. Tell me if the numbers are too small to trust. Use only the details I give you here. If something you need is missing, write [missing] and tell me what to add, rather than making it up.
Watch outA handful of hires a quarter is too small a sample for firm answers. Treat the output as questions to look into.
Draft AI ground rules for your team
Use whenYour recruiters already use AI and nobody has written down what is allowed.
Draft a one page guide to using AI in recruiting for a [team size] team at a [type of company] in [country or state]. Cover: which candidate data must never go into a personal or unapproved AI account; which outputs a person must check before they are used; which decisions AI must never make alone, such as rejecting a candidate; what we tell candidates about AI in our process; and what we keep on record. Write it for recruiters, not lawyers: plain language, short bullet points, and one line on who to ask when unsure. If you are not sure something is accurate or current, mark it (check) rather than stating it as fact.
Watch outThis is a first draft to take to HR or legal, not legal advice. Rules differ by state and country.
Compare your desk’s numbers with typical benchmarks
Use whenYou want to know whether a number is a real problem or normal for your kind of desk.
Here are my recruiting numbers for [quarter or year]: [paste, e.g. time to fill, submittals per interview, interviews per offer, offer acceptance rate, placement fall offs, revenue per recruiter, average fee]. I run [contingent, retained, RPO or in house] recruiting for [niche]. For each number, say whether it looks strong, typical or weak for this kind of desk, explain what that view rests on and label it as an estimate. Where I look weak, suggest one specific process change and the number that would show it worked. Tell me which benchmarks I should confirm from a published survey.
Watch outAI benchmarks are rough and often out of date. Check them against a trade body survey or your own past quarters before you act.
Find your most profitable role types
Use whenYou are busy all quarter, but the fees do not match the hours.
Here are my placements for [period]. Each row has role title, industry, fee, days to fill, hours I spent and whether the person is still in the job: [paste]. Work out the fee per hour for each placement, showing the sums. Then rank role types by return on effort, and flag any placements that did not last long enough to be worth it. Tell me where my time is out of line with the fee, and where I may be charging too little for the work. If my hours are estimates, say how much that weakens the answer. Use only the details I give you here. If something you need is missing, write [missing] and tell me what to add, rather than making it up.
Watch outHours are usually guessed after the fact. Log time on your next few searches before you drop a role type.
Summarize a market map for the client
Use whenThe research is done and the client wants to know what the market looks like before they see names.
Here are the findings from my market map for a [role] search: [number of people and companies reviewed, how many are strong fits, where they sit now, pay ranges seen, where the talent is concentrated, what is making the search hard]. Write a summary of about 300 words for [the CEO, the board or the hiring manager] that: 1) describes the talent pool in plain terms; 2) picks out the three findings that matter most for this search; 3) explains the search approach those findings point to; 4) sets honest expectations on timeline and obstacles; 5) ends with the decisions I need from them. Use only my figures and do not name any candidate.
Watch outCheck every figure against your research notes. Clients remember the pay range you quoted.
How to get better results
What should the weekly update say?
Where the search stands, what changed, what is blocking it and what you need from them. Write the weekly hiring manager update builds that from your notes in the same order every week.
How do you find what slows hiring down?
Paste the count at each stage and the average days in each from your ATS into Find where your hiring process stalls. Ask for the working so you can check the math.
Do you need team rules for AI?
Yes, before someone pastes a resume into a personal account. Draft AI ground rules for your team gives you a one page starting point to agree together.
Keep going
Quick answers
3 questionsCan ChatGPT analyze recruiting data?
Yes. ChatGPT, Claude, Grok and Gemini can read a spreadsheet or ATS export, work out conversion rates and time in stage and draw charts. Ask for the working and spot check a few numbers yourself.
What recruiting metrics should I report weekly?
For each open role: candidates at each stage, what moved since last week, time open and the one thing blocking progress. Keep it the same every week so the hiring manager knows where to look.
Is it safe to upload an ATS export to ChatGPT?
Only to an account your company approves. Exports carry names and contact details, so remove those columns first if you can.
More recruiting prompts
Learn the rest in the AI Academy.
Eight free modules that take a recruiter from first prompt to automations that run on a schedule.