A pressure testing prompt system for job interviews, discovery calls, and client onboarding.
This guide assumes you have never used an AI prompt, prepared a discovery call, or run client onboarding.
A prompt is simply a set of instructions you paste into an AI chat. The prompt tells the AI how to behave. In this kit, the AI becomes a demanding but useful coach.
The coach does not help by saying everything sounds great. It helps by finding what would break in the real conversation.
Before talking about yourself, understand the person on the other side.
This is empathy engineering: not guessing feelings, but studying the other side’s operational reality and preparing around it.
The flow is:
The coach labels information:
Use this when preparing for a recruiter, hiring manager, panel, or final interview.
The coach studies the company and role, identifies likely business or team pain, and tests whether your real experience connects to it.
Minimum inputs:
| Input | Example | Why it matters |
|---|---|---|
| Company | Northstar SaaS | Gives the story a real target |
| Role | Customer Success Manager | Defines the expected responsibility |
| Source | Job description or link | Provides evidence for the pain |
| Your evidence | CV or one named experience | Prevents invented achievements |
| Interview context | Hiring manager interview | Changes the likely follow-ups |
Use this before a first client, sales, qualification, advisory, or discovery call.
The coach prepares pain hypotheses and questions. It does not pretend the prospect has confirmed the diagnosis.
Minimum inputs:
| Input | Example | Why it matters |
|---|---|---|
| Prospect | Brightline Agency | Gives the call a real buyer |
| Source | Website and call notes | Grounds the hypotheses |
| What they deliver | Marketing campaigns for local brands | Shows where work may leak |
| Buyer | Agency founder | Identifies whose pressure matters |
| Your call goal | Earn a paid workflow audit | Keeps the conversation purposeful |
| Your proof | Reduced reporting time in a prior project | Prevents generic pitching |
Use this after the client says yes, signs, or agrees to work together.
This mode prepares delivery, not sales. It looks for scope creep, missing stakeholders, absent access, unclear acceptance, bad handoff, unrealistic timelines, and missing ownership.
Minimum inputs:
| Input | Example | Why it matters |
|---|---|---|
| Client or project | Brightline reporting workflow | Names the delivery context |
| Agreement | Proposal or email summary | Establishes what is actually agreed |
| Expected outcome | Weekly report produced in two hours | Gives the project a measurable purpose |
| Stakeholders | Founder and account lead | Exposes authority and ownership |
| Biggest unknown | CRM access | Finds the first delivery risk |
CALL-READINESS-COACH.md.Do not ask the coach to give you every question at once. The next question must depend on the answer you just gave.
Priority is fixed:
OVERSHARE > OFF-TOPIC > WEAK OWNERSHIP > BUZZWORD > RAMBLING / TOO LONG > GOOD
| Flag | What it means | Simple example | Why it matters |
|---|---|---|---|
| OVERSHARE | You exposed damaging, confidential, desperate, or irrelevant information | “I hated my manager and needed the money.” | Protects trust and leverage |
| OFF-TOPIC | You answered a different, easier question | Asked “why?” but answered with the result | The real question remains unanswered |
| WEAK OWNERSHIP | You hid inside “we”, “the team”, “they”, or passive language | “We decided to change onboarding.” | The listener needs your contribution |
| BUZZWORD | Your words sound polished but explain nothing | “We leveraged cross-functional alignment.” | Empty language fails under follow-up |
| RAMBLING / TOO LONG | The useful point is buried | Two hundred words before the answer | The listener may miss the evidence |
| GOOD | The answer is direct, concrete, relevant, and owned | “I replaced the handoff checklist and measured stalled accounts weekly.” | The drill can advance safely |
OFF-TOPIC, WEAK OWNERSHIP, BUZZWORD, and RAMBLING always block the drill. OVERSHARE wins priority. It may advance only if the useful answer remains complete after the damaging detail is removed.
OVERSHARE covers bad-mouthing or blaming a manager, colleague, or
employer, confidential personnel detail such as who was fired or quit,
desperation or pay complaints, and any damaging personal commentary
about another person. Damaging talk about someone else is OVERSHARE, not
WEAK OWNERSHIP. Simply naming another person does not make an answer
WEAK OWNERSHIP. If you already said clearly what you personally did, the
coach removes the damaging detail and advances; it does not relabel a
complete, owned answer as WEAK OWNERSHIP. Leading with the damaging
detail does not force a retry either: if your remaining answer still
holds the substance and your own action, the coach cuts the oversharing
part for you and moves on with Next: rather than making you
tell the whole story again.
An unisolated “we”, “our team”, or passive collective triggers WEAK OWNERSHIP even when the underlying claim is usable. The coach never awards GOOD until you isolate your personal action.
Every coach turn after one of your answers begins with the flag, then takes one of two shapes:
Blocking (OFF-TOPIC, WEAK OWNERSHIP, BUZZWORD, RAMBLING / TOO LONG, and incomplete OVERSHARE) keeps you on the same question:
[FLAG: TAG] Keep <useful part>; cut <exact defect>.
Retry: <the same question, narrower>Advancing (GOOD, and a complete OVERSHARE after the cut) moves you to the next question:
[FLAG: GOOD] Keep <useful part>; cut <minor surplus if any>.
Next: <one question for the next answer target>If you see Retry:, answer that same question more
tightly. If you see Next:, you passed and the drill moves
on.
User input
Job Interview Mode. I am interviewing for Customer Success Manager at Northstar SaaS. The job description owns onboarding, adoption, renewals, and reporting. My claim is that I managed 45 accounts and reduced time-to-value.
Coach question
Walk me through the onboarding change in order, just the spine.
Bad answer
We aligned stakeholders and improved the onboarding journey with better reporting.
Coaching flag
[FLAG: WEAK OWNERSHIP] Keep the onboarding topic; cut “we” and the group-level claim.
Narrower retry
What did you personally change first?
Improved answer
I mapped the handoff from sales, found that new accounts had no named kickoff owner, and added a mandatory owner and kickoff date before an account could move into implementation.
Why pressure
Why did naming the owner matter more than adding another training session?
Improved deeper answer
Because the delay was not lack of knowledge. Nobody was accountable for moving the customer from signed contract to first value. Training without ownership would leave the same gap.
Northstar’s role is responsible for onboarding, adoption, and renewals, so I would lead with a case where I fixed the ownership gap between sales and customer success. I mapped the handoff across 45 accounts and found that kickoff delays happened because no one owned the transition. My first attempt was an extra enablement session, but it did not change the delay. I then added a named kickoff owner, required a date before implementation, and reviewed stalled accounts weekly. That reduced time-to-value because the process had clear accountability, not because we added more activity. Without the original CRM setup, I could rebuild the operating checklist, but I would first stall on validating the historical baseline.
Expected follow-up whys:
Why was ownership the root cause?
How did you know your change caused the improvement?
What would you measure at Northstar in the first 30 days?
Ready answers: the delay happened before product entry; compare kickoff delay before and after the ownership gate; measure time to kickoff, time to first value, and stalled-account age.
Say this: “I changed the handoff rule and named the owner.”
Never say this: “We aligned stakeholders and transformed onboarding.”
Readiness verdict: Ready on ownership and mechanism; prepare stronger baseline evidence.
User input
Prospect Discovery Mode. Brightline is a 15-person marketing agency. Their site shows a growing client roster. My contact said reporting takes too long and handoff varies. I want the next step to be a paid workflow audit. I previously reduced weekly reporting from six hours to two.
Coach question
Which pain is supported by a source rather than only guessed?
Bad answer
They clearly need an AI transformation to optimize cross-functional delivery.
Coaching flag
[FLAG: BUZZWORD] Keep the idea that delivery may be inefficient; cut “AI transformation” and “cross-functional optimization.”
Narrower retry
What concrete pain did the contact actually report?
Improved answer
The contact said weekly client reporting takes too long and that account handoff differs by manager. CRM hygiene is still a hypothesis I need to test.
Coach question
What question would test the reporting pain without leading the buyer?
Improved answer
Walk me through how last week’s client report was assembled from the moment work ended to the moment the client received it.
From what you shared, the reported issue is the time required to assemble weekly reports, while CRM hygiene and its contribution to handoff inconsistency still need validation. I would not start by proposing a tool. I would first map one real report from source data to client delivery, identify where manual reconciliation or unclear ownership creates delay, and quantify the cost. In a prior project, that approach reduced weekly reporting preparation from six hours to two. If the same pattern exists here, the useful next step would be a paid workflow audit. If it does not, I would say so rather than force the offer.
Expected follow-up whys:
Why audit before automation?
What evidence would justify changing the workflow?
What would make you recommend no project?
Ready answers: the audit prevents pricing the wrong intervention; inspect one real report, its sources, owners, time, and errors; recommend no project when the delay is rare or low-cost.
Say this: “CRM hygiene is a hypothesis I need to test.”
Never say this: “You obviously need an AI transformation.”
Readiness verdict: Ready to diagnose; do not present unconfirmed CRM pain as fact.
Brightline has agreed to a reporting workflow project. The proposal says “automate weekly client reporting” but does not name a decision owner, acceptance test, CRM access owner, or change rule.
User input
Client Onboarding Mode. Brightline agreed to a four-week reporting workflow project. The goal is to reduce reporting preparation from six hours to two. The founder and account lead are known. CRM access and acceptance criteria are unclear.
Coach question
What does the proposal say the client will accept at the end?
Bad answer
The project is complete when everyone is happy and the workflow is streamlined.
Coaching flag
[FLAG: BUZZWORD] Keep the need for client acceptance; cut “everyone is happy” and “streamlined.”
Narrower retry
What observable condition lets the client accept the deliverable?
Improved answer
UNKNOWN - ASK ON CALL. My proposed test is that the workflow produces the agreed report for five sample clients, the account lead verifies the numbers, and the founder signs the acceptance checklist.
Coach question
Who owns the final acceptance decision?
Improved answer
UNKNOWN - ASK ON CALL. The founder is a hypothesis, but the onboarding call must name the owner.
Prep Brief: [USER] The user reports a four-week project and a six-to-two-hour target. [SOURCE] The proposal lacks acceptance and change control. [USER] Founder and account lead are known. [HYPOTHESIS] The founder may approve. [USER] CRM access is a known risk.
Client objective: [USER] the user reports an intended reduction from six hours to two.
Stakeholder map:
Scope: [SOURCE] map and implement the weekly reporting workflow named in the proposal.
Out of scope: UNKNOWN - ASK ON CALL. No exclusions were supplied.
Access needed: [USER] CRM access is a named risk. Access owner and method are UNKNOWN - ASK ON CALL. [USER] The user proposed five sample clients for acceptance; client approval and access are UNKNOWN.
Inputs and dependencies: UNKNOWN - ASK ON CALL. No dependency owners or delivery deadlines were supplied.
Timeline and milestones: [USER] the user reports that four weeks is agreed; detailed review dates are UNKNOWN - ASK ON CALL.
Acceptance criteria: [USER] The user proposes reports for five sample clients, verified figures, and a signed acceptance checklist. Client approval remains UNKNOWN - ASK ON CALL.
Definition of done: UNKNOWN - ASK ON CALL. The user supplied a proposed acceptance test, but not a complete definition of done.
Change-request rule: UNKNOWN - ASK ON CALL. No rule was supplied.
Risks, open questions, and Naked Test stall: acceptance authority, access owner, data quality, reliable CRM access, and field ownership.
Questions to ask first:
Who has final acceptance authority?
Who grants and maintains CRM access?
Which five reports define the acceptance sample?
Say this: “That field is not agreed yet, so I am marking it UNKNOWN until this call assigns it.”
Never say this: “We will figure out scope as we go.”
Readiness verdict: Not ready to start delivery until acceptance authority and CRM access have owners.
Job and Prospect modes produce:
Client Onboarding Mode produces:
The public kit is provider-neutral. It works from pasted sources and, when available, ordinary web research. It never needs access to someone else’s machine.
The private operator kit may connect the same Stage 0 workflow to specialist research helpers. Those helpers can speed up company, prospect, person, fit, and premortem analysis, but they do not change the evidence rule: sources are facts, user claims are claims, and inferences are hypotheses.
| Mistake | Correction | Why it matters |
|---|---|---|
| Pasting the guide instead of the prompt | Paste CALL-READINESS-COACH.md |
The guide explains; the prompt operates |
| Asking for all questions at once | Let the coach adapt one turn at a time | Later probes depend on the prior answer |
| Arguing with a negative flag | Answer the narrower retry | The flag must change behavior |
| Treating web research as internal company truth | Label it SOURCE or HYPOTHESIS accurately | Public evidence is not private certainty |
| Inventing a clean onboarding scope | Mark missing fields UNKNOWN - ASK ON CALL | Prevents accidental commitments |
| Rehearsing silently | Speak the sharpened answer aloud | Spoken delivery exposes weak phrasing |
This Markdown guide and its branded HTML version are prepared for review. No PDF has been generated. After approval, the HTML can be rendered to PDF using the same Systems Detective navy, gold, and ivory visual system.