Designs an employee referral program for one role or team, with the ask to staff built from the scorecard, a guide to what makes a good referral, bonus rules as placeholders, the same assessment for referred people and a thank-you and update loop. Use for "run recruit-employee-referral-program", "employee referral program", "referral bonus rules", "ask staff for referrals", "referral email to employees", "how to get better referrals", "referral program policy", "warm introductions instead of ...
Installs into .claude/skills of the current project.
Are you the author of Recruit Employee Referral Program?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/polar-bear-org-recruit-employee-referral-program)
---
name: recruit-employee-referral-program
description: Designs an employee referral program for one role or team, with the ask to staff built from the scorecard, a guide to what makes a good referral, bonus rules as placeholders, the same assessment for referred people and a thank-you and update loop. Use for "run recruit-employee-referral-program", "employee referral program", "referral bonus rules", "ask staff for referrals", "referral email to employees", "how to get better referrals", "referral program policy", "warm introductions instead of inbound", part of the AI for Recruiting Pack by Polar Bear.
---
# Employee Referral Program
## When To Use
Inbound is full of fakes and pasted CVs, and you want warm introductions from people who have seen the work, without the hire narrowing to people who look like the current team. It answers: what do we ask staff for, how do we treat what they send, and how do we keep referrers coming back?
## When Not To Use
If you are choosing the whole channel mix for a role (outbound, boards, communities, past applicants), run Sourcing Strategy; this designs one channel in depth. If staff are unhappy enough that they would not refer a friend, fix that first; no bonus buys a referral from someone planning to leave.
## Inputs
- The Job Scorecard: outcomes and must-haves.
- Your current referral rules, if any: bonus, eligibility, timing, who runs it.
- The Hiring Process Plan stages, so referred people follow the same path.
If you have none of this, I start from the role title and its must-haves, leave every rule as a [placeholder], and mark the output as a first draft.
## Approach
Referral programme design as SHRM sets it out (shrm.org, "How to Create an Employee Referral Program"): a clear ask, simple rules, and referrers who hear back. The EEOC (eeoc.gov, "Prohibited Employment Policies/Practices") warns that word-of-mouth recruiting alone can keep a workforce the way it already is, so referrals run beside other channels, never instead of them. The failure it prevents: "know anyone good?" in the all-staff channel, twenty friends of friends, none matching the job, and referrers who never learn what happened and stop sending anyone.
## Workflow
1. Ask at most three questions: which roles the program covers, whether a bonus exists and who approves it, and which other channels run for this role at the same time.
2. Write the ask from the scorecard: the outcomes and the three to five must-haves, in words staff would use at lunch. Never "know anyone good?".
3. Write the "good referral" guide: what the referrer saw the person do, in what setting, and confirmation they asked the person first. A name alone is not a referral.
4. Set the bonus rules as [placeholders] the user fills: amount, who is eligible (hiring managers for their own role usually are not), when it pays, and whether part is tied to the hire staying [n] months. Check with a qualified adviser on tax and local rules.
5. State the same-path rule: referred candidates go through the same stages and questions as everyone else, and a referral skips nothing.
6. Build the update loop: the referrer hears at each stage (received, in process, closed) within [days], with no assessment detail, and a thank-you whatever the outcome.
7. Add the mix check: referrals are one channel beside the others from the Sourcing Strategy, reviewed after [n] weeks for whether the pool is narrowing.
## Output Format
```markdown
# Employee Referral Program: [Role or team]
## The ask to staff
[Message: the outcomes, the must-haves, how to refer, by when.]
## What makes a good referral
- [What you saw them do, where] | [You asked them first] | [How they match a must-have]
## Bonus rules
| Rule | Setting | Approved by |
|---|---|---|
| Amount | [amount] | [role] |
| Eligibility | [who can and cannot claim] | [role] |
| Timing | [on start / after [n] months] | [role] |
## Same path for referred candidates
[Stages and questions, identical to every other applicant.]
## Update loop
| Stage | Message to referrer | Sent within |
|---|---|---|
| Received | [message] | [days] |
| Closed | [thank-you, no reasons given] | [days] |
## Decision
[Talent acquisition lead] approves the bonus rules and the channel mix check date by [date], then sends the ask.
```
## Done When
- The ask names the outcomes and must-haves, not a vague request.
- Every bonus figure and period is a [placeholder] with an approver.
- The same-path rule is written into the program, in plain words.
- The update loop gives referrers status only, never notes or reasons.
## Quality Bar
- A good referral describes work the referrer saw; friendship alone does not qualify.
- Referrers never see interview notes or the reasons for a decision.
- Referrals run beside other channels, and the program says when the mix is reviewed.
- Referred people meet the same questions; Claude never rates a referral.
## Next
Run recruit-boolean-search (Boolean Search Strings) to reach the people no one on staff knows yet.
## About the makers
This pack is made by Polar Bear, a consultancy built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).