How to Use This Prompt Pack
Each prompt below names its job, the phase of the coaching cycle it serves, and the bracketed placeholders you replace before running it. Two disciplines make the pack work. The first is the context rule: prompts that prepare and structure get rich context, role, relationship history, the development area, one anonymized paragraph about the situation. Prompts that analyze real conversations get summaries, never names. The second is the adaptation rule: AI drafts, you decide. Every output is raw material for a conversation between two humans, and a manager who pastes model output verbatim into a session is not coaching, a coachee can smell a script from three questions away.
The twelve prompts follow the arc of a real development relationship: architect the session, open the conversation, convert insight into commitment, then follow through. If you want the framework underneath before the prompts, the 15-minute G.R.O.W.S. method explains why each phase exists and where managers typically lose the thread. These free coaching prompts assume that structure and automate the paperwork around it.
Why Prompt Discipline Beats Coaching Software
The market is full of coaching platforms, and most of them are bottlenecks wearing dashboards. The leverage in coaching does not sit in software; it sits in the quality of the questions and the discipline of the follow-up. The published numbers make the case plainly. Zenger Folkman's survey, published by Harvard Business Review, found 92 percent of respondents agreeing that corrective feedback, delivered well, improves performance, and 57 percent actively preferring it to praise. Yet the Interact/Harris poll found 69 percent of managers uncomfortable communicating with the people they manage, and Gallup found only 7 percent of US workers strongly agree that communication at work is accurate, timely, and open.
That is a demand-supply gap: near-universal appetite for honest development conversations, near-total scarcity of them. A prompt library attacks the supply side directly, because it lowers the cost of preparation until preparation actually happens. The workplace coaching statistics behind that pattern are worth reading in full before you blame busy calendars for thin conversations, the gap is a skills-and-systems problem, not a scheduling one.
Sources: Zenger Folkman survey, Harvard Business Review 2014 · Interact/Harris Poll 2015 · Gallup, State of the American Workplace 2017
Phase One: Architect the Session (Prompts 1-3)
Preparation is where coaching conversations are won or lost, and it is the phase AI helps with most safely, because nothing here touches the live human. Run these prompts the morning of the conversation, not in the hallway outside the room.
Prompt 1: the G.R.O.W.S. session architect
Prompt: "Act as my coaching preparation partner. I coach [ROLE] on [DEVELOPMENT AREA]. Design a 15-minute G.R.O.W.S. session plan: two Goal questions that produce a specific target, three Reality questions that surface facts without blame, an Options sequence, one Will question that yields a single commitment, and a Support checklist. Flag the three points where managers typically drift into advice-giving." The drift flags are the real deliverable, you are rehearsing the failure modes, not just the questions. A plan that names the temptation to advise is far easier to resist at minute nine than a vague intention to listen.
Prompts 2-3: the question bank and the listening audit
Prompt 2: "Generate 10 open coaching questions for [TOPIC]: three that clarify the goal, three that examine current reality, two that expand options, two that test commitment. No why-questions that sound like blame. Keep each question under 12 words." The word cap matters; short questions invite answers, long ones invite performances.
Prompt 3 is the uncomfortable one, run it on yourself: "Here is what I remember saying in my last coaching conversation: [SUMMARY]. Audit me: estimate my ratio of questions to statements, flag leading or stacked questions, mark where I likely interrupted, and offer two better questions for each moment." Managers who run this listening audit quarterly report the same finding: they talk more than they think, and the excess talk is usually advice.
Phase Two: Open the Conversation (Prompts 4-6)
The opening minutes decide whether the session belongs to the coachee or to you. These three prompts cover the three most common openings that fail: the vague goal, the correction disguised as development, and the pushback nobody planned for.
Prompt 4: the goal sharpener
Prompt: "My coachee described their goal as [VAGUE GOAL]. Give me the five questions most likely to turn it into a specific, observable target, then show what the sharpened version sounds like in one sentence the coachee could own as their own words." The final clause is the important one: a goal phrased in the manager's language produces compliance, while a goal the coachee could repeat in their own words produces ownership. Those are different outcomes that happen to use the same whiteboard.
Prompts 5-6: the feedforward drafter and the resistance navigator
Prompt 5 builds on Marshall Goldsmith's feedforward method: "Draft a 10-minute feedforward exchange for [SKILL]: two future-focused suggestions I offer, then the invitation script asking for their suggestions to me. No review language, no references to past mistakes." The reversal step, inviting suggestions about your own behavior, is what separates feedforward from rebranded advice. The SBI+ feedback model remains the better tool when a specific past incident must be examined; feedforward is for momentum, not adjudication.
Prompt 6 handles the moment the conversation stalls: "My coachee responded to feedback with [DESCRIBE: silence, deflection, over-agreement]. Give me three responses that preserve psychological safety and return ownership to the coachee, plus one question I should NOT ask, and why." Rehearsing pushback before it happens follows the same logic as AI rehearsal for hard conversations, the first live attempt should never be the first attempt.
Phase Three: Convert Insight Into Commitment (Prompts 7-9)
Insight without commitment is entertainment. The closing phase turns a good conversation into a plan with an owner, a deadline, and a safety net, and three prompts do the drafting so the live minutes stay focused on judgment rather than paperwork.
Prompt 7, the options expander: "Here is the situation, anonymized: [ONE PARAGRAPH]. Generate 8 options: two safe, two stretch, two borrowed from another industry, two that cost nothing. One trade-off line each. Do not rank them, selection is the coachee's job." The no-ranking instruction protects ownership at the exact moment managers most want to grab the steering wheel. Prompt 8, the will planner: "Turn this option, [SELECTED OPTION], into a commitment plan: a first action under 30 minutes, a deadline, a definition of done, and the two most likely failure points with pre-planned responses. One page maximum." Prompt 9, the support map: "For that commitment, list what makes it stick: what I do as manager, what the coachee controls, what the organization must supply, and what we deliberately will not do. Two columns: Support and Owner." The will-not-do column is where scope creep goes to die.
Phase Four: Follow Through (Prompts 10-12)
What happens between sessions is where development actually compounds, and it is the phase nobody schedules. These three prompts make follow-through light enough to survive a real calendar.
Prompt 10, the session record: "Turn these rough notes, [PASTE], into a five-line record: focus, key insight in the coachee's words, commitment, support, and the opening question for next time. First-person plural, nothing identifying." Five lines is the threshold that matters: records long enough to require formatting never get written twice. Prompt 11, the progress review: "Design a 10-minute review of this commitment, [PASTE]: two questions that test evidence rather than intention, one that honors progress, and decision rules, continue, adjust, or escalate, for each answer pattern." Prompt 12, the coach self-audit, run quarterly: "Act as my coaching supervisor. Here are my last three session records, anonymized: [PASTE]. Diagnose my patterns: am I coaching or consulting, where do I close exploration too early, and what is the one habit to change next quarter?" The best managers treat that prompt the way athletes treat film review.
The Evidence Base: What Coaching Conversations Control
If the pack needs a business case, the engagement research supplies it. Gallup attributes roughly 70 percent of the variance in team engagement to the manager, which makes every manager conversation an engagement instrument. The same research program found that 80 percent of employees who say they received meaningful feedback in the past week are fully engaged, while only one worker in three strongly agrees they received recognition or praise in the last seven days. At the team level, Gallup's Q12 meta-analysis ties top-quartile engagement to 23 percent higher profitability than the bottom quartile.
The retention line is equally direct: LinkedIn Learning's Workplace Learning Report found 94 percent of employees would stay longer at a company that invests in their development. Conversations are that investment in its cheapest and most frequent form, which is why a disciplined prompt library is a retention tool wearing a coaching costume. The numbers below are the argument in one view.
Sources: Gallup, State of the American Workplace 2017 · Gallup, 2022 · Gallup · LinkedIn Workplace Learning Report 2018 · Gallup Q12 Meta-Analysis 2020
The Privacy Firewall: Three Rules
Coaching content is among the most sensitive material a manager handles, and prompts make it portable. The firewall is three rules, written here so you can paste them into your team's shared document. First, anonymize by default: roles and situations, not names, identifying project details, or anything touching health, performance documentation, or legal exposure. Second, tier your tools: consumer chat tiers for frameworks and question banks; your approved business tier with training opt-outs for anything resembling a real case; never any tier for what your compliance team would call a record. Third, outputs are drafts: an AI-suggested plan about a real person gets reviewed, edited, and owned by you before it influences anything.
- Replace names with roles and industries before pasting anything
- Business tier with training opt-out for case-like material; consumer tier for frameworks only
- If you would not email it externally, do not paste it anywhere
| Material | Tier allowed | Rule of thumb |
|---|---|---|
| Frameworks, question banks, session structures | Any tier, including free consumer chat | No person involved yet, so nothing to leak |
| Anonymized role-and-situation summaries | Business tier with training opt-out | Pattern described, name absent, still case-like enough to tier up |
| Real session records and five-line notes | Business tier with opt-out, stripped first | First-person plural, nothing identifying |
| Health, performance documentation, legal exposure | No tier, ever | If compliance would call it a record, it never touches a model |
Source: Compiled from ZLearning Coaching and Mentoring Conversations privacy guidance 2026
Common Failure Modes and Their Fixes
Four failures account for most disappointments with AI-assisted coaching, and each has a one-line fix.
- Generic output, your context line was too thin; add stage, relationship history, and the last conversation's focus
- Advice-flavored questions, add the constraint: no suggestions phrased as questions
- Overlong session plans, cap the plan at 15 minutes and five questions, or the structure eats the dialogue
- Prompt fatigue, you do not need all twelve; prompts 1, 4, and 8 carry most of the value in week one
What Good Looks Like in 30, 60, and 90 Days
Adopting the pack is a calendar change, not a tooling change. Thirty days: run prompts 1, 4, and 8 before every significant one-to-one, and write the five-line session record within an hour of each conversation. The observable outcome is boring and important, every conversation ends with a first action and a date, and nobody leaves wondering what was decided.
Sixty days: add the resistance navigator and the progress review, and run the listening audit on one reconstructed conversation. Ninety days: run the coach self-audit against three session records, pick the single habit to change, and ask each coachee one calibration question, what did our last conversation change about your week? Structured practice pairs naturally with this timeline; the free coaching course day one linked below is designed for exactly that rhythm.
| Checkpoint | What you run | Observable outcome |
|---|---|---|
| Day 30 | Prompts 1, 4, and 8 before every significant one-to-one, plus the five-line record | Every conversation ends with a first action and a date |
| Day 60 | Add the resistance navigator and progress review; one listening audit on a reconstructed conversation | Stalled commitments get re-opened instead of quietly dropped |
| Day 90 | Coach self-audit against three records; one calibration question per coachee | One named habit to change next quarter, chosen from evidence |
Source: Compiled from ZLearning Coaching and Mentoring Conversations adoption timeline 2026
The Complete Coaching System
These prompts are the skeleton; the Coaching & Mentoring Conversations program is the musculature, the full question banks, demonstration sessions, and practice loops that turn good prompts into fluent coaching. Day 1 is free with any account, and it hands you a session plan and a question set you can use the same afternoon.
For managers who want the credential as well as the craft, the Leadership Coaching Mastery certification formalizes the skill set for internal coaching roles. Start with the prompts, adopt the two disciplines, and let the evidence do the persuading: the conversation layer is the cheapest performance lever you control.
Frequently asked questions
▸Are these prompts safe to use with real coaching situations?
Yes, if you follow the privacy firewall: anonymize roles and situations, use business-tier tools with training opt-outs for anything case-like, and keep performance, health, and legal matters out of consumer chat tiers entirely. The prompts are written to work on anonymized descriptions, if one fails on a role-and-industry summary, the problem is usually missing context, not the redaction. When in doubt, describe the pattern, not the person.
▸Which three prompts should I start with?
Prompts 1, 4, and 8, the G.R.O.W.S. session architect, the goal sharpener, and the will planner. Together they cover the three moments that decide a coaching conversation: the plan before it, the target inside it, and the commitment after it. Add the listening audit in month two and the coach self-audit at the end of the quarter; the rest earn their place as specific situations demand them.
▸Do these prompts work with free versions of ChatGPT, Claude, or Gemini?
Yes, every prompt runs on any current chat model, free or paid. Paid tiers help with long-context work such as auditing three session records at once, but question banks, session plans, and feedforward drafts produce their full value on free tiers. The tool matters far less than the two disciplines: specific context in, human adaptation out.
▸Can AI prompts replace actual coaching skills?
No, and a prompt that tried would be a liability. The prompts automate preparation and follow-through, the parts of coaching that die in busy calendars, so the live minutes go to listening, noticing, and asking one better question. Managers who treat outputs as scripts get formulaic sessions; managers who treat them as rehearsal keep their flexibility when the conversation goes somewhere unplanned.
▸How is feedforward different from feedback?
Feedback analyzes a past event; feedforward generates suggestions for a future one. Goldsmith's method replaces the post-mortem, which reliably triggers defensiveness, with a two-way exchange of future-focused suggestions, and it works because nothing is on trial. Use the SBI+ model when a specific incident must be examined, and feedforward when the goal is momentum.
▸How often should managers run these coaching prompts?
The preparation prompts (1, 4, and 8) run before every significant development conversation, typically weekly or biweekly per direct report. The listening audit and the coach self-audit are quarterly instruments; running them weekly turns reflection into rumination. If the pack feels heavy, cut prompts before you cut session records, because the five-line record is the habit that compounds.
▸Where can I learn the coaching frameworks behind these prompts?
Day 1 of the Coaching & Mentoring Conversations program is free with any account and covers the G.R.O.W.S. model, question design, and feedforward in a working-session format. To judge any free program against a standard before spending your evening, the free coaching course evaluation guide shows exactly what a real first day should deliver.