HR AI Prompts: Job Descriptions, Interviews and Performance Reviews

A workforce plan compares the work the business needs done with the people it has, and decides for each gap whether to build (develop someone internally), buy (hire) or borrow (contract).

• Build takes 12 months or more: both customer success team leads have under two years in role and have never owned a renewal target. Developing either would leave the gap open through every renewal in the next year.
• Borrow costs $20K a month and owns nothing: the interim consultant covering since August runs the weekly meeting but cannot hire, reset targets or own the renewal plan. Four months of cover adds $80K to the $207K already lost on salary, fees and severance.
• Buy protects $11.5M: 96 accounts worth $11.5M renew in the next six months. At a 97% net revenue retention rate, each point of retention on that book is worth $115K.

Job analysis is the study of what a role must achieve and which skills produce that result; the job description is written from it. Skills-based requirements state what a candidate must be able to do and show, instead of the years or qualifications they hold.

• The old description described a person: “10+ years in SaaS” and “strategic leader” were the two lines the last hire met best. He left after 7 months, and the mis-hire cost $287K.
• Four outcomes replace the adjectives: net revenue retention back to 105%, attrition under 15%, the $11.5M renewal book protected, and a named successor within 18 months. Each has a baseline and a date.
• One requirement does the work of three: “Has rebuilt a customer success team after losses, and can show the retention figures” replaced years, an MBA and a title, because it is the only one of the four that describes the situation Orrin is in.

An employer value proposition (EVP) is the set of reasons a person would choose to work for a company, stated as what the job gives them. A sourcing strategy decides where candidates will come from and how many must be approached to make one hire.

• The claim no employee made: “fast-paced, collaborative culture” appeared in the old advert and in none of 70 employee survey comments. It was cut and replaced with what the job actually is.
• The difficulty is the selling point: the advert states that the team lost 3 of its 9 customer success managers and needs rebuilding. That excludes candidates looking for a stable team, which is the profile that failed last time.
• 40 approaches, and a $51K fee avoided: at Orrin’s past conversion rates, 40 targeted approaches produce about 14 CVs and one offer. An agency at 25% of the $205K base would cost $51K, so it is used only if there is no shortlist by week 5.

A structured interview asks every candidate the same questions in the same order and rates each answer on an anchored scale, where each score point is defined by an observable behaviour.

• Four interviews, no shared question: on the last hire, no two interviewers asked the same question, so there was nothing to compare. All four rated him highly on impressions, and the $287K cost followed.
• The anchor decides the score: a 1 describes the team’s effort in general terms; a 5 names the metric, the cause found and the result. Two interviewers hearing the same answer now give the same score.
• 9 of 14 CVs out on two knockouts: leading a team of 8 or more and owning a retention number. Each rejection records the missing evidence, so the decision can be explained to the candidate.

A selection decision combines interviewers’ scores into a hire or no-hire recommendation. Scores are written independently before the debrief, because once a senior person states a view, other interviewers’ scores move towards it.

• 3.4 against 4.2: written before the debrief, the panel’s scores put Candidate B ahead on five of six competencies. The CEO’s preference for A rested on presentation, which the scorecard does not measure.
• A 2-point spread is a question, not an average: two interviewers scored A’s team leadership 2 and 4. The debrief found one had asked about a team of 3, the other about a team of 12.
• References tested a figure: B’s claim of cutting churn from 14% to 8% was confirmed by a former CFO. A’s claimed retention result could not be confirmed by either referee.

A pay percentile shows where a salary sits in the market: at the 75th percentile, 75% of comparable roles pay less. Internal equity compares an offer with what peers in similar roles are paid.

• $215K breaks two limits: it is above the 75th percentile of $214K and above the internal band maximum of $210K, where two peer heads of function sit at $195K and $208K.
• $41K paid only for results: 20% variable on the $205K base is $41K, paid against the net revenue retention targets in the job description, so the extra pay is earned through retention.
• The walk-away was written first: the limit of $210K base and the items that could move before it, such as a signing bonus and a six-month review, were agreed with the CFO before the offer call.

A 30-60-90 plan sets what a new hire learns, diagnoses and delivers in each of their first three months. Time to productivity is the time until the hire performs the role’s core work without support, and it is measured against criteria written before day one.

• Two accounts, $0.6M, before the first customer meeting: the last hire spent month one in product training. Two accounts worth $0.6M did not renew before he met them in month three.
• 41% of revenue met in 30 days: the top 20 accounts hold $17.2M of the $42M, so meeting them first covers the largest share of risk in the fewest meetings.
• A deliverable, not a feeling, at 90 days: the day-90 review asks one question: is there a renewal plan for the $11.5M due by March, with an owner for every account?

Calibration is a meeting in which managers compare their draft ratings against the same standard before any rating is shared. SBI (Situation, Behaviour, Impact) is a way of writing feedback: where it happened, what the person did, and what it changed.

• 7 of 9 “exceeds” from one manager: the customer success team lead rated 7 of 9 people “exceeds”; the support lead rated 3 of 11. The same results were being scored on two different scales.
• 2 of 3 leavers named it: two of the three customer success managers who resigned said in exit interviews that there was no difference between strong and weak performers.
• “Great attitude” is not evidence: rewritten as SBI, it became: at the March renewal with a top-20 account, found the unused module, and the account renewed with a 12% increase.

A 9-box grid places each person on current performance and on potential, the evidence they can handle a larger role. The 70-20-10 model plans development as roughly 70% experience on the job, 20% learning from others and 10% formal training.

• No successor, $287K exposure: if the new head of customer success left in year one, the role would go back to an external search and interim cover. The last time, that cost $287K.
• Potential rated on scope, not presence: both team leads have held their roles under two years. The one given the stretch assignment has already run a renewal cohort of $1.1M without escalation.
• 70% is the assignment: owning the $3.2M Q3 renewal cohort, with the head of customer success as backstop, tests the skill the succession gap needs. A course alone would not.

Retention risk analysis compares the people who left with the people who stayed, to find the factors that predict leaving. A stay interview asks a current employee what would make them leave and what would make them stay.

• 11 of 14 leavers: had no pay review in the 12 months before they resigned. Exit surveys most often gave “career opportunity” as the reason.
• 4 people, $124K: four of the 11 support staff have had no pay review in 18 months. Replacing one costs about half a $62K salary, $31K, so the four together cost $124K.
• 31 of 70 comments: named no career path. Stay interviews start with the strongest performers in support, and the pay reviews are brought forward to next month.