Stop typing what you could say in 10 seconds.

Wispr Flow turns your voice into clean, professional text inside any app. Emails, Slack, client updates — speak once, send without editing. 4x faster than typing.

How frequently are you getting referrals from your customers? You've got clients who would say yes if you asked them the right way. Most of them have never been asked at all.

I started tracking where my best clients actually came from instead of assuming, and referrals outperformed every channel I was deliberately working.

And that's the part reps get backwards. A satisfied customer and an advocate are not the same person. Your NPS score and your check-in calls measure satisfaction. They don't tell you who's actually willing to spend their own social capital vouching for you to someone else, because those are two different questions and most processes only ever ask one of them.

This isn't the closed-won lookalike workflow from S01E27. That mines your best accounts for firmographic and signal patterns to find similar new prospects. This is about the customers sitting in your book right now, and getting the right ones to open a door you can't open yourself.

Three moves fix this: find the right handful of customers, time the ask to a moment that's actually earned it, and make the ask specific enough that they don't have to do the matching work themselves.

Why the Ask Doesn't Work

Most reps ask for referrals once, at the wrong time, with the wrong question.

The wrong time is a quarterly calendar reminder or the tail end of a renewal call, tacked on after the real conversation is already over. The wrong question is "know anyone else who could use us?" That question works against you twice. It asks a satisfied customer to do research on your behalf, scanning their own network for a fit they were never asked to think about. And it gives them an easy out, because "nothing comes to mind right now" costs them nothing to say.

So most reps ask once, get a shrug, and conclude their customers don't refer. The customers were never given a real chance to.

Advocates Aren't the Same as Satisfied Customers

A survey response costs nothing. A referral costs the customer something real: their reputation with someone they know, spent on vouching for you. That's why a referral is a far stronger signal than a satisfaction score, and why most companies are measuring the cheap signal instead of the expensive one.

Before you ask anyone for anything, you need to know who's already shown you they're willing to spend that capital, because they usually tell you before you ask.

The Referral Engine

Step 1: Find the Right Handful Not your whole book. Look for customers who have already referenced you unprompted: mentioned you in a Slack community, forwarded your content to someone, left an unprompted compliment on a call, renewed early, or expanded without a pitch. Each of those is proof of advocacy you didn't have to ask for. That's your real candidate list, usually far shorter than your customer list.

Step 2: Time the Ask to a Moment They Just Earned Not a calendar reminder. Ask right after a moment of proven value: right after a renewal, right after a QBR where a real result got named out loud, right after they tell you something worked. The ask lands because the value is still fresh, not because it's the first of the quarter.

Step 3: Make the Ask Specific, Not Generic "Know anyone who could use us" makes them do the work. A specific ask does the work for them: name a role, a company type, or a situation close to their own. Better still, if you can see a name in their network worth asking about directly, ask about that name.

Why did one company's AI work, and another's didn't?

One had a dedicated owner. Resolution rate: 48.9%. One didn't: 0.38%. See the full breakdown.

The Advocate Extraction Prompt

Run this against your renewal notes, QBR notes, and any unprompted customer messages you've saved:

You are extracting voice-of-customer signal from the full post-sale
relationship, not just the most recent call.

From the attached notes and transcripts, covering onboarding,
implementation, check-ins, and any upsell or expansion conversations,
find every instance where the customer said something positive and
specific, not just "good," "happy," or "no complaints."

For each instance found, extract:
- The customer and their role
- Which stage it happened in: onboarding, implementation, check-in,
  upsell, or other
- The specific thing they said, quoted or closely paraphrased
- What kind of praise it is: a measurable result (time, cost, revenue,
  headcount), a comparison to something they used before, or an
  experience ("easier than I expected")
- A specificity score, 1 to 3: 1 means vague sentiment with nothing to
  build on. 2 means a real result, but thin on detail. 3 means specific
  and quotable as-is.

Then, for each instance:
- If it scored 1 or 2, draft one follow-up question the rep can ask on
  their next call to turn it into something usable
- If it scored 3, draft two things: a specific referral ask built from
  that exact language, and a short, low-pressure testimonial request
  asking permission to use the quote

Do not include generic positivity unless it's attached to a specific
reason or result. "Really happy with it" is not signal. "Cut our
turnaround from three days to one" is.

A Worked Example

Illustrative: a mid-market ops manager who renewed two months early and told you on the call that the rollout "saved her team a full day a week."

Run the prompt and it surfaces her as a real signal, not just a happy renewal: early renewal plus a specific, quotable result. The drafted ask isn't "know anyone else who could use us." It's built off her own words: a specific ask about anyone she knows running ops at a similar-sized team, sent the same week as the renewal, while the result is still the thing she'd tell someone about anyway.

Why This Works

The filter works because it separates a real signal from a cheap one. Satisfaction is easy to give and easy to forget. A referral costs something, so the customers who've already shown you they're willing to spend it are the only ones worth asking first.

The timing works because the ask rides on a moment that already proved the value, instead of asking someone to remember why they liked you three months ago.

And the specific ask works because it removes the one thing that kills most referral requests: making the customer do research they never agreed to do.

Getting Started

Pull your renewal and QBR notes from the last two quarters. Run the Advocate Prompt against them this week.

Don't ask everyone who comes back with a positive mention. Start with the two or three who show real advocacy signal, not just satisfaction, and send the specific ask while the moment is still fresh.

~ Jay

PS: Today’s issue finds the advocate and drafts your ask to them. For paid subscribers: below is the follow-up prompt that drafts the actual intro message your advocate sends on your behalf, the one that does the real work once they've said yes.

Click here for the tool stack I’d build today: https://www.sellingwithai.vip/stack

Author curated previous issues:

Click here for the entire archive: https://www.sellingwithai.vip/archive

Members Section

The Warm Intro Prompt

Once an advocate has agreed to make an introduction, most of them stall out because they don't know what to write. Draft it for them instead of leaving it to chance:

logo

You just read the motion. Now run it.

The prompts, checklists, and templates that turn this into a 10-minute execution are in the Vault.

Get Vault Access

Vault access includes::

  • Copy and paste execution prompt packs
  • Deal, outbound, and follow-up playbooks
  • Operating checklists for every motion
  • Members Vault access

Keep Reading