Referral Program Copilot
Referrals become a system with a memory, not a thing you remember to do twice a year
The problem
For most service businesses, referrals are the best channel and the least managed one. The [Referral Request Scripts](/toolbox/referral-request-scripts) write one good ask, and then the system ends: nobody tracks who was asked, who said "of course!" and never sent anyone, which thank-you went out, or that your best referrer has now sent three clients and received nothing but a text. Word of mouth runs on timing and reciprocity, and both need a memory. This skill installs it.
The tool
Once installed, this skill turns your assistant into the keeper of your referral program. It learns your business once (who your best clients are, what a good referral looks like, what you offer as thanks, if anything) and then runs the standing loop: who is ask-ready this month, the ask drafted in your voice with the forwardable blurb included, and the log updated from your two-line reports.
The memory is the machine. Ask-ready means something specific: a client who got a result recently, praised it, and has not been asked in months; the skill watches for that window from what you report and surfaces names at the right moment instead of in your annual guilt sweep. It tracks every ask and its outcome, so nobody gets re-asked awkwardly and quiet yeses get one graceful nudge. It keeps the thank-you ledger, flagging the referrer owed a gesture before the imbalance gets embarrassing. And it reads the pattern: when four referrals in a row came from one industry, it says so and suggests where the next asks should concentrate.
What's inside
- A one-time setup: your clients' shape, the referral you want, your thank-you practice
- A monthly ask-ready list with the reason each name is on it
- Asks drafted in your voice, forwardable blurb included, one nudge maximum per yes
- The full log: asked, said yes, sent someone, thanked, gone quiet
- A thank-you ledger that flags what you owe before the relationship notices
- Pattern reads: where referrals concentrate, and where to aim next
How to use it
- 1Click "Install skill" on this page and copy the install prompt.
- 2Paste it into your AI assistant (Claude, ChatGPT, anything that accepts pasted text).
- 3Do the setup, then seed the log with the referrals you can remember from the last year.
- 4Report in two lines as things happen ("asked Dana, she said yes", "new client via Marcus").
- 5Once a month, ask for the ask-ready list and send what it drafts, edited to taste.
Example
A bookkeeping practice seeds the skill with its client list and last year's four remembered referrals. Month one, the ask-ready list has three names, each with a reason ("year-end cleanup praised two weeks ago, never asked"). Two asks go out, one client says yes. Month two's report includes "new client signed, came via Priya," and the skill answers with the log updated and a flag: Priya has now referred twice and the thank-you ledger shows nothing sent, with a suggested gesture matched to the setup answers. By month four the pattern note reads: three of five referrals are restaurant owners; the skill suggests the next round of asks lead with the two restaurant clients not yet asked, and drafts a blurb that names the niche.
Pro tip
Seed the log honestly, including the asks that went nowhere; the skill's "one graceful nudge, then stop" rule only protects relationships if it knows who was already asked. And when the pattern read names a niche, act on it in your marketing too; a referral concentration is your market telling you where you already have a reputation, which is intelligence worth more than the referral itself.
Related tools
Want this customized and automated for your business?
We take the tools in this toolbox and wire them into your business: your data, your brand voice, running on autopilot.
Talk to Vexlo