Muse for Sales: Prospecting and Account Research on Autopilot

Muse does the research and the first drafts, and you make the calls and send the messages. Pre-call briefs, prospecting lists, and post-meeting follow-ups, with an approval gate on everything that leaves your desk.

Jolly the cream-furred creature at a warm-lit sales desk with a whiteboard of prospect notes behind it, soft-3D, no text

Sales is two jobs wearing one title. The first is research: who is this company, who decides, what do they care about, what did we already discuss. The second is follow-up: the email after the call, the nudge after the silence, the next step that keeps a deal from going cold. Both jobs are mostly reading and writing, which is exactly the work Muse is good at. The part Muse never does is the part that closes deals: the call itself, the judgment call on a discount, the message that goes out under your name.

That is the line for this whole lesson. Muse preps, humans send. Every draft below ends with you reading it before anything leaves your desk, because Meta’s rule for Muse is that nothing publishes, sends, or spends without your approval. The sales workflow I run respects that rule and builds on it.

What Muse Plugs Into

On September 29, 2026, Meta announced Muse for Small Business, which brought connectors into the same Muse app: HighLevel, Slack, Notion, Zoom, Klaviyo, QuickBooks, Shopify, Stripe, and more, plus Facebook Pages, Instagram professional accounts, and Meta ad accounts. Three of those connectors carry the load for sales work. HighLevel holds leads, conversations, and calendars. Email and calendar hold the history of who you have talked to and what you promised. Meeting notes, from Zoom or a notes tool, hold what was actually said on the last call.

If your pipeline lives in a CRM that is not on the connector list, you still have options. Muse supports custom connectors, and a regular export of your pipeline into a spreadsheet Muse can read covers the same ground for a one-person operation. The setup below assumes HighLevel because it is the connected CRM, but the prompts work against any pipeline Muse can see.

The Setup: The Pipeline Desk

The Pipeline Desk is the name I give the whole arrangement, and I run it for a solo rep selling a Shopify app to independent retailers. One pipeline, one rule: Muse reads everything, drafts everything, sends nothing. Every weekday morning, a scheduled routine runs three jobs before the first call. It builds the day’s pre-call briefs, flags the deals that have gone quiet, and drafts the follow-ups from yesterday’s calls. The rep starts the day reading a page of briefs instead of opening twelve tabs.

The Pipeline Desk has three jobs, and each one maps to a prompt below.

Pre-call briefs. Before every call, Muse assembles one page: what the company does, recent public news about it, who else at the company you have spoken to, what came up last time, and the open items. A good brief means you walk into the call knowing the context instead of spending the first ten minutes re-establishing it. The product manager version of this same move is the morning triage; the sales version just points it at accounts instead of a roadmap.

Prospecting lists. Give Muse the shape of a good customer and it builds a list with a reason attached to every name. The reason matters more than the name. A list of fifty companies with no reason to call them is a phone book. A list of twenty with one concrete, sourced reason each is a week of real calls.

Post-meeting follow-ups. After a call, Muse reads the notes and drafts the follow-up: thanks, restates what the prospect asked for, confirms the next step and the date. You read it, fix anything it got wrong, and send it. Speed matters here more than anywhere else in sales. A follow-up that arrives the same afternoon reads as competence. A follow-up that arrives Thursday reads as an afterthought.

Job Muse does You do
Pre-call brief Researches the account, assembles one page, cites sources Reads it, corrects it, makes the call
Prospecting list Finds companies matching your criteria, attaches one sourced reason each Picks the list, works the phones
Post-meeting follow-up Drafts from the call notes, restates asks and next steps Edits, approves, sends

The specialist pattern applies here as it does everywhere else: one setup drafts, one setup researches, and no setup sends. If you share a pipeline with teammates, keep each person’s Muse scoped to their own accounts and their own memory, the same way the customer support lesson keeps drafting and triage apart. And keep the privacy rules in mind: prospect data that is public is fair game for research, but a prospect’s confidential information, pricing you quoted in confidence, and anything under NDA stays out of prompts.

The short version

  • Muse preps, humans send. Muse researches accounts, builds lists, and drafts follow-ups. You make the calls and send the messages.
  • Every pre-call brief gets one page: what the company does, recent news, prior conversations, open items, and questions worth asking.
  • A prospecting list is only as good as the reason attached to each name. No sourced reason, no call.
  • Follow-ups go out the same day. Muse drafts from the call notes, you edit and send.
  • Nothing sends without your approval. That is Meta’s rule for Muse, and it is the rule that makes this whole setup safe to run.

The Copy-Paste Prompts

Prompt 1: The pre-call brief

Run this the evening before, or the morning of, for each call on the calendar:

I have a call tomorrow at [time] with [name] at [company]. Build me a one-page pre-call brief: what the company does, recent public news about it, who else at the company I have spoken to and what we discussed, the open items from our last conversation, and three questions worth asking. Cite the source for every fact. If you cannot find a source for a fact, leave it out and say what is missing. Do not invent anything.

Prompt 2: The prospecting list

Build a prospecting list of [number] [type of company] in [region] that [criteria, e.g. “sell on Shopify and have under 20 employees”]. For each one, give me: company name, website, the most likely decision-maker title, and one concrete reason they fit, with a source link. Skip any company where you cannot find a source for the fit. Rank the list by strength of fit.

Prompt 3: The post-meeting follow-up

Read the notes from my call today with [name] at [company]. Draft a follow-up email: thank them, restate the two things they asked for, confirm the next step and the date we agreed, and name anything I owe them. Keep it under 150 words. Draft only. I send it.

What Can Go Wrong: The Confident Prospect

The failure mode you will meet first is the brief that sounds perfect and is wrong. Muse hands you a pre-call brief saying your prospect just raised a Series B and is hiring aggressively. You open the call by congratulating them. The funding round was two years ago, the hiring page is stale, and the prospect now wonders what else you got wrong. Nothing in the brief flagged the mismatch, because the agent sounds confident whether the facts check out or not. If you are reading five briefs on a busy morning, approval fatigue is doing its work, and the wrong fact goes into a live conversation under your name.

The fix is in Prompt 1 and it is non-negotiable: every fact gets a cited source, and anything without a source gets left out with the gap named. That single rule turns a fluent guess into a checkable claim. The second habit is a thirty-second check before each call: click the two most load-bearing links in the brief, the funding claim and the news claim, and confirm they say what the brief says they say. Most wrong briefs trace back to a stale article or a misread page, not a bad agent. And remember what the Sentinel approvals gate: Sentinel gates actions Muse takes, like sending. It does not fact-check a brief. The verification is your job, every time.

Jolly the cream-furred creature working a warm-lit service counter with glowing chat bubbles floating above, soft-3D, no text Role guides Muse for Customer Support: Working the Queue

Sales run this way replaces the part of the job that was never judgment: the tab-hopping research, the blank-page follow-up, the prospect list with no reason attached. Your judgment stays yours. The call is yours, the discount call is yours, the send button is yours. And the prospect who gets a same-day follow-up that restates exactly what they asked for knows they are dealing with someone who was listening. That is the whole edge. The edge is showing up prepared, on time, every time.

“Muse preps, humans send.”

Keep reading

  • Jolly the cream-furred creature at a warm-lit law library desk with stacked contract folders and a reading lamp, soft-3D, no text

    Muse for Legal Work: Review Prep Without the Risky Bits

    Muse organizes contracts, preps counsel meetings, and triages a legal inbox, while the lawyer keeps every judgment call. First-pass summaries, hearing prep, and inbox triage with a verification gate on everything, plus the ethics rules that set the boundaries.

  • Jolly the cream-furred creature at a small founder's desk with a laptop, a wall of sticky notes, and a coffee mug, warm morning light, soft-3D, no text

    Muse for Founders: Running a One-Person Company

    A founder's operating setup for Muse: investor updates that write themselves from real numbers, a hiring pipeline that never goes stale, competitor watching that stays quiet until it matters, and a Monday prep routine that starts the week with the desk already cleared.

  • Jolly the cream-furred creature working a warm-lit service counter with glowing chat bubbles floating above, soft-3D, no text

    Muse for Customer Support: Working the Queue

    Staff the front door with Meta Business Agent and work the queue with Muse: morning triage, draft replies a human sends, a knowledge base that stops rotting, and escalation rules that hold.