Capstone — The Full Employee Lifecycle

Ayesha Raza joined the EdTech company in Karachi eighteen months ago. She was hired as a Senior Data Analyst into Omar Farooq's team. She has been through an onboarding programme, her first performance review, a compensation benchmarking exercise, and now — as the company grows — a succession conversation about a leadership role that has opened up. And if, hypothetically, Ayesha were to leave two years from now, the offboarding and knowledge capture process would make sure the knowledge she accumulated did not leave with her.

Ayesha's journey is what the tools in this chapter were built for. Not individual isolated use-cases — a JD here, a performance review there — but a coherent, end-to-end workflow that uses the right tool at the right moment in a continuous relationship between an employee and an organisation.

This capstone lesson asks you to run that lifecycle yourself. You will play the role of the HR team supporting Ayesha's journey — or an employee from your own organisation, anonymised if needed. You will use every skill and agent in the chapter stack. And at the end, you will reflect on what the AI did well, where it needed human judgment, and what you would configure differently in hr.local.md if you were doing it again.

The Skills and Agents in This Chapter

Before starting the lifecycle sprint, here is the complete map of what you have built across L01-L13.

Official Plugin Skills (Anthropic human-resources)

Command	What It Does	Lifecycle Stage
/policy-lookup	Synthesise policies into plain-language summaries	Continuous (L03)
/onboarding	Generate 30-60-90 onboarding plan	Stage 2: ONBOARD
/draft-offer	Draft offer letters and employment documents	Stage 1: HIRE
/interview-prep	Prepare interview questions and evaluation rubrics	Stage 1: HIRE
/performance-review	Structure and draft performance reviews	Stage 3: DEVELOP
/comp-analysis	Benchmark compensation at percentile bands	Stage 1: HIRE, Stage 3: DEVELOP
/org-planning	Model org structure changes	Stage 4: RETAIN/PROMOTE
/people-report	Generate headcount, attrition, diversity reports	Stage 6: CONTINUOUS
/recruiting-pipeline	Track and diagnose the recruiting funnel	Stage 1: HIRE

Custom Plugin Skills (Panaversity hr-operations)

Command	What It Does	Lifecycle Stage
/jd	Write job descriptions with inclusive language	Stage 1: HIRE
/match	Assess internal candidates for succession	Stage 4: RETAIN/PROMOTE
/knowledge	Capture institutional knowledge	Stage 5: OFFBOARD
/reference	Draft reference letters and employment verification	Stage 1–5 (on request)
/offboard	Structure the offboarding process	Stage 5: OFFBOARD

Persistent Agents (Panaversity hr-operations)

Agent	What It Does	When It Runs
knowledge-base-agent	Answers employee HR questions 24/7	Always-on
onboarding-orchestrator	Runs pre-boarding through Day 90 workflow	Triggered by HRIS new hire record
policy-maintenance-agent	Monitors policy currency and statutory rates	Monthly + on rate changes
offboarding-knowledge-agent	Automates knowledge capture on resignation	Triggered by HRIS resignation record

The Employee Lifecycle: Six Stages

The lifecycle follows Ayesha from before she joined to — hypothetically — the day she might leave. Each stage maps to the skills and agents you have already learned.

Stage 1: HIRE

Skills: /jd, /interview-prep, /comp-analysis, /draft-offer Agents running in background: knowledge-base-agent (answers candidate HR questions via recruiting channel if configured)

Before Ayesha could join, someone had to write a compelling job description for the Senior Data Analyst role. Someone had to prepare structured interview questions. Someone had to benchmark the salary to make sure the offer was competitive. And someone had to draft an offer letter that was accurate for Pakistan jurisdiction and reflected the company's actual benefits.

/jd Senior Data Analyst — EdTech company, Karachi, Pakistan

Department: Analytics (reports to Head of Analytics)
Level: Senior
Key responsibilities: [brief description]
Required skills: [list]
Nice-to-have: [list]
Inclusive language required: yes
Jurisdiction: Pakistan

Then:

/interview-prep Senior Data Analyst — EdTech, Karachi

Role: Senior Data Analyst
Level: Senior
Interview format: 3-stage (recruiter screen, technical, hiring manager)
Key competencies to assess: analytical thinking, technical skills (SQL, Python,
data visualisation), stakeholder communication, learning mindset
Jurisdiction: Pakistan

Then:

/comp-analysis Senior Data Analyst — Karachi, Pakistan

Company: EdTech, ~250 employees. Role is IC, no direct reports.
Market: Pakistan technology sector.
Provide percentile bands for base salary. Note internal equity consideration:
existing Data Analysts in team are at 150,000-170,000 PKR/month.

Finally:

/draft-offer Ayesha Raza — Senior Data Analyst

Company: EdTech company, Karachi, Pakistan
Jurisdiction: Pakistan (Employment Ordinance 1968, applicable provincial regulations)
Candidate: Ayesha Raza
Role: Senior Data Analyst
Start date: [Date]
Salary: 190,000 PKR/month
Probation: 3 months
Benefits: [list from hr.local.md]
Reporting to: Omar Farooq, Head of Analytics

What runs automatically: The onboarding-orchestrator triggers when Ayesha's new hire record appears in the HRIS. It begins the pre-boarding sequence: IT provisioning, system access requests, welcome email, pre-boarding task list.

Sensitivity: /jd output = ROUTINE. /interview-prep output = ROUTINE. /comp-analysis output = CONFIDENTIAL. /draft-offer output = CONFIDENTIAL.

Stage 2: ONBOARD

Skills: /onboarding Agents running in background: onboarding-orchestrator (pre-boarding → Day 1 → Day 30 → Day 60 → Day 90), knowledge-base-agent (Ayesha's Day 1 questions)

/onboarding Ayesha Raza — Senior Data Analyst

Role: Senior Data Analyst
Manager: Omar Farooq, Head of Analytics
Start date: [Date]
Team context: 3-person analytics team; EdTech product company
Key priorities in first 90 days: [list]
Jurisdiction: Pakistan
Tools/systems to access: [list]
30-day goal: [specific outcome]
60-day goal: [specific outcome]
90-day goal: [specific outcome]

The /onboarding skill produces a structured 30-60-90 plan with milestones, check-in points, and success criteria for each phase. The onboarding-orchestrator runs the logistics — task checklists, survey sends, manager reminder messages — automatically.

On Day 1, Ayesha has a question about how to book leave. She asks the knowledge-base-agent. It answers immediately with a plain-language summary and a link to the policy source. She does not need to message Sarah in HR.

Sensitivity: /onboarding output = ROUTINE (plan is not performance-sensitive; it describes the process, not the person).

Stage 3: DEVELOP

Skills: /performance-review, /comp-analysis Agents running in background: knowledge-base-agent, policy-maintenance-agent

After twelve months, Omar and Ayesha have her first formal performance review. Omar uses /performance-review to structure his thinking and draft the review document.

/performance-review Ayesha Raza — Year 1 Review

Role: Senior Data Analyst
Review period: 12 months
Manager: Omar Farooq
Company: EdTech, Karachi, Pakistan
Jurisdiction: Pakistan

Performance evidence:
- Led the customer analytics dashboard project (delivered on time, strong stakeholder feedback)
- Improved reporting pipeline efficiency (reduced run time from 4 hours to 45 minutes)
- Mentored two junior analysts on SQL optimisation techniques
- Identified data quality issue in revenue reporting — raised proactively

Development areas:
- External stakeholder presentations: capable but needs confidence at leadership level
- Strategic thinking: good at execution; developing ability to set the analytics agenda

Compensation review: salary at 190,000 PKR/month, below 50th percentile after 12 months
of inflation. Review appropriate.

After the review, Omar uses /comp-analysis to benchmark Ayesha's salary for the upcoming pay review — checking whether 190,000 PKR/month is still competitive after a year of market movement.

Sensitivity: /performance-review output = CONFIDENTIAL. /comp-analysis output = CONFIDENTIAL.

Stage 4: RETAIN/PROMOTE

Skills: /match, /org-planning, /comp-analysis Agents running in background: knowledge-base-agent

Eighteen months in, a Team Lead, Analytics opportunity opens. Omar wants to assess Ayesha as an internal candidate — but also has Bilal Ahmed in the picture.

This is the succession planning workflow from L09. Omar runs /match for both candidates, reviews the six-dimension output, has the succession conversation with Ayesha using the conditional pathway language ("if your trajectory continues, a leadership role becomes realistic"), and uses /org-planning to model what the analytics team looks like if Ayesha moves up.

Sensitivity: /match output = CONFIDENTIAL. /org-planning output = ROUTINE (structural model). Succession conversation notes = CONFIDENTIAL.

Stage 5: OFFBOARD (if applicable)

Skills: /offboard, /knowledge Agents running in background: offboarding-knowledge-agent (triggered by resignation record)

Three years later — hypothetically — Ayesha decides to pursue an opportunity elsewhere. She gives notice. The offboarding-knowledge-agent triggers immediately, generating a risk-calibrated knowledge capture plan based on her role, tenure, and the knowledge assets she holds.

HR uses /offboard to structure the process:

/offboard Ayesha Raza — Senior Data Analyst (now Team Lead Analytics)

Jurisdiction: Pakistan
Last day: [Date] (4 weeks notice)
Risk level: HIGH (leads analytics team; holds key stakeholder relationships;
  deep knowledge of revenue reporting architecture)
Systems to revoke: [list]
Knowledge at risk: customer analytics dashboard, revenue reporting pipeline,
  three key client analytics relationships
Handover: to [successor name or "TBD"]

And uses /knowledge to structure a knowledge capture interview before Ayesha leaves:

/knowledge Ayesha Raza — knowledge capture interview guide

Role: Team Lead Analytics (formerly Senior Data Analyst)
Tenure: 3 years
Exit date: [Date]
Time available: 3 sessions of 90 minutes each
Knowledge at risk:
  - Revenue reporting architecture (she designed it; not fully documented)
  - Customer analytics dashboard (she built it; team relies on her for edge cases)
  - Key client relationships: [list]
  - Analytics team processes (daily standup norms, escalation paths, tool access)
Output: structured knowledge base articles for each domain

What runs automatically: The offboarding-knowledge-agent sends the knowledge capture interview schedule, creates handover task checklists for Ayesha and her manager, and monitors completion.

Sensitivity: /offboard output = CONFIDENTIAL. /knowledge output = CONFIDENTIAL. Handover plan = CONFIDENTIAL.

Stage 6: CONTINUOUS

Skills: /people-report, /policy-lookup Agents running in background: All four, continuously

Throughout Ayesha's three years, the four persistent agents have been running:

• knowledge-base-agent answered her questions every time she had one — about leave, about benefits, about the promotion process
• onboarding-orchestrator ran her onboarding and will run for every hire after her
• policy-maintenance-agent ensured that the policies she read on Day 1 remained accurate as statutory rates changed
• offboarding-knowledge-agent is now capturing everything she knows before she leaves

And the CHRO has been running /people-report quarterly, seeing Ayesha's first year attrition cohort, tracking the team's engagement scores, watching for signals in the KB agent weekly report that point to issues before they become problems.

The Lifecycle Sprint Exercise

Type: Capstone Practice Time: 90 minutes Plugin commands: All Goal: Produce a complete employee lifecycle folder for one employee — real (anonymised) or fictional — using the correct skill for each stage

This exercise is one continuous workflow, not six isolated tasks. Work through the stages in order. The output of each stage informs the next.

Setup (5 minutes)

Choose your employee:

• Option A: Use Ayesha Raza and the EdTech company (all details are in this lesson)
• Option B: Use a real employee from your organisation (anonymise the name and any identifying details)
• Option C: Create a fictional employee for a fictional company you define

Define:

• Employee name (or alias)
• Role and level
• Manager name
• Company: size, sector, location, jurisdiction
• Key career events: hire date, first review date, any promotion or succession event, exit date (real or hypothetical)

Stage 1: HIRE (20 minutes)

Produce three documents:

1. Job description — Use /jd
2. Interview plan — Use /interview-prep
3. Offer letter — Use /draft-offer (benchmark salary first with /comp-analysis if you want to set a realistic figure)

Label each: ROUTINE or CONFIDENTIAL.

Stage 2: ONBOARD (15 minutes)

Produce one document:

4. 30-60-90 onboarding plan — Use /onboarding

Verify: does it include observable success criteria for each phase? Does it specify who is responsible for each milestone? Label: ROUTINE.

Stage 3: DEVELOP (15 minutes)

Produce one document:

5. Performance review — Use /performance-review

Include: performance evidence, development areas, at least one specific competency rating with a rationale. Label: CONFIDENTIAL.

Stage 4: RETAIN/PROMOTE (15 minutes)

Produce one document:

6. Talent assessment — Use /match to assess your employee against a leadership or next-level role

Include: six-dimension assessment, readiness classification, development plan if applicable. Draft the opening line of the succession conversation you would have (what you would say, using the conditional pathway language). Label: CONFIDENTIAL.

Stage 5: OFFBOARD (10 minutes)

Produce two documents:

7. Offboarding plan — Use /offboard
8. Knowledge capture interview guide — Use /knowledge

Identify the top three knowledge risks (what this employee knows that is not documented elsewhere). Label both: CONFIDENTIAL.

Deliverable

A lifecycle folder containing:

• 8 documents (JD, interview plan, offer letter, onboarding plan, performance review, talent assessment, offboarding plan, knowledge capture guide)
• Each document with the correct sensitivity label
• A one-paragraph reflection per stage: what did the AI do well, and where did it need human judgment?

Keep This File

The lifecycle folder you build here is a portfolio artefact — it demonstrates that you can operate an AI-native HR function end to end. It is also the reference point for Lesson 15's quick reference tables, which map every skill and agent to the stage where it appears in this lifecycle.

The Human Judgment Reflection

After completing the lifecycle sprint, answer these questions for each stage:

Stage	What the AI did well	Where human judgment was required
HIRE
ONBOARD
DEVELOP
RETAIN/PROMOTE
OFFBOARD
CONTINUOUS

Guiding questions for the reflection:

• HIRE: Did the offer letter require editing for your specific jurisdiction? What did the JD get right about inclusive language, and what did it miss?
• ONBOARD: Did the 30-60-90 plan include observable success criteria, or did it list activities without outcomes? Which milestones required manager knowledge that the AI could not have?
• DEVELOP: How much editing did the performance review require? Where did the structured format help, and where did it produce generic output that only you could make specific?
• RETAIN/PROMOTE: What did the /match six-dimension framework surface that a gut-feel assessment might have missed? What did it miss that gut-feel might have caught?
• OFFBOARD: Did /offboard correctly identify all the knowledge risks? What tacit knowledge does your employee hold that a tool could not have known to ask about without your input?
• CONTINUOUS: What would the CHRO need to monitor about this employee's journey that the agent reports would not surface automatically?

Configuration Reflection: What Would You Add to hr.local.md?

Based on what you observed during the lifecycle sprint, identify two or three things you would add to hr.local.md to improve the quality of future outputs:

## hr.local.md improvements I would make

1. [Configuration gap observed]: [What I would add]
   Example: "The /draft-offer output used generic notice period language.
   I would add: notice_period_probation: "1 week" and
   notice_period_post_probation: "4 weeks" to the Pakistan section."

2. [Configuration gap observed]: [What I would add]

3. [Configuration gap observed]: [What I would add]

The Complete Skill and Agent Map

Every skill and agent in the chapter, mapped to where it appeared in this lifecycle:

Tool	Type	Stage(s)	Lesson Taught
/policy-lookup	Official skill	Continuous	L03
knowledge-base-agent	Custom agent	Continuous (always-on)	L04
/onboarding	Official skill	Stage 2: ONBOARD	L05
onboarding-orchestrator	Custom agent	Stage 2: ONBOARD	L05, L12
/jd	Custom skill	Stage 1: HIRE	L06
/interview-prep	Official skill	Stage 1: HIRE	L06
/draft-offer	Official skill	Stage 1: HIRE	L07
/reference	Custom skill	Stage 1-5 (on request)	L07
/performance-review	Official skill	Stage 3: DEVELOP	L08
/comp-analysis	Official skill	Stage 1: HIRE, Stage 3: DEVELOP	L09
/match	Custom skill	Stage 4: RETAIN/PROMOTE	L09
/org-planning	Official skill	Stage 4: RETAIN/PROMOTE	L09
/knowledge	Custom skill	Stage 5: OFFBOARD	L10
/offboard	Custom skill	Stage 5: OFFBOARD	L11
offboarding-knowledge-agent	Custom agent	Stage 5: OFFBOARD	L11, L12
policy-maintenance-agent	Custom agent	Continuous (monthly)	L12
/people-report	Official skill	Stage 6: CONTINUOUS	L13
/recruiting-pipeline	Official skill	Stage 1: HIRE	L13

This table is what an AI-native HR operation looks like: 9 official skills + 5 custom skills + 4 persistent agents, each in the right place, each handling the right part of the lifecycle. Not a replacement for HR judgment — a multiplier of it.

Try With AI

Try With AI

Use these prompts in Cowork or your preferred AI assistant.

Reproduce: Run the full lifecycle sprint for Ayesha Raza using the scenario in this lesson.

I am going to run a full employee lifecycle exercise for one employee.
I will work through six stages and produce a document at each stage.

Employee: Ayesha Raza
Role: Senior Data Analyst (joining), later Team Lead Analytics (after promotion)
Manager: Omar Farooq, Head of Analytics
Company: EdTech company, ~250 people, Karachi, Pakistan
Jurisdiction: Pakistan

Start with Stage 1: HIRE.

Step 1: Write a job description for the Senior Data Analyst role.
Requirements: 3+ years experience, strong SQL and Python, experience with
BI tools (Tableau or Power BI), data engineering familiarity an asset.
Use inclusive language. Pakistan market.

Label the output with its sensitivity level (ROUTINE or CONFIDENTIAL).
Then pause and wait for me to confirm before moving to the next step.

What you are learning: Working stage by stage with a pause for review between each step teaches you to treat each skill output as a draft that requires human judgment before proceeding — not a finished document.

Adapt: Map the lifecycle to your own organisation's most recent hire.

I want to map the AI-native HR lifecycle to a recent hire in my organisation.

Employee: [Name or role — anonymise if needed]
Role: [Title and level]
Manager: [Name or role]
Company: [Brief description: size, sector, jurisdiction]
Timeline: Hired [date], current status [e.g., "12 months in, first review due"]

For each lifecycle stage below, tell me:
1. Which skill or agent I should use
2. What inputs it would need
3. What the primary human judgment moment is at that stage

Stages: HIRE, ONBOARD, DEVELOP, RETAIN/PROMOTE, OFFBOARD, CONTINUOUS

What you are learning: Mapping the lifecycle framework to a real hire makes abstract tools concrete. The "primary human judgment moment" question is the most valuable part — it reveals where the AI is a tool that supports a decision versus the decision itself.

Apply: Design your organisation's AI-native HR playbook.

I want to design an AI-native HR playbook for my organisation.

Organisation context:
  - Size: [headcount]
  - Sector: [sector]
  - Jurisdiction(s): [primary locations]
  - HR team size: [number]
  - Current HR tools: [HRIS, ATS, etc.]

Using the hr-operations and human-resources plugin stack, design:
1. Which lifecycle stages are the highest priority to implement first (and why)
2. What hr.local.md configuration is needed for my jurisdiction
3. Which skills the HR team should learn in the first month
4. Which persistent agents to deploy first and in what sequence
5. What the HR team's workflow looks like 90 days after deployment

Be specific about what "AI-native HR" looks like for my organisation's size and context.

What you are learning: Designing a playbook for your own organisation forces you to sequence the implementation — which is harder than running all tools at once. Priority reveals what you actually believe about where the value is.

Flashcards Study Aid

---

Continue to Lesson 15: Quick Reference & Central Insights →