AI prompts for sales managers do a different job than the rep-facing lists that dominate search results. A frontline manager is not writing cold emails; you are coaching reps from call recordings, interrogating a deal list before the Monday meeting, building a forecast narrative your VP will actually read, and walking into eight 1:1s a week with something better than "how's the pipeline looking." Each of those is a repeatable task with inputs you already have — transcripts, CRM exports, rep metrics — which makes each one promptable. The prompts below cover the manager workflow specifically, with the paste-this-first step that makes the output usable.
What managers actually use AI for
The pattern across working sales managers is consistent: the value is not in generating words, it is in structured review of material you do not have time to read closely. A 45-minute call transcript, a 60-row pipeline export, a quarter of rep activity data — these are exactly the inputs a model handles well when the prompt tells it what standard to apply. The manager prompts that stick are review prompts, not writing prompts.
Coaching prompts: from call transcript to scored review
"Here is a call transcript: [paste]. Score the rep against this standard: did they establish the problem before presenting product, did they quantify impact, did they secure a concrete next step, did they talk less than 45% of the time. For each item, quote the moment from the transcript that supports your score, then give one specific behavior to change on the next call. Do not comment on anything outside these four items."
The scoring rubric is the point. Without one, the model produces vague encouragement; with one, it produces evidence-quoted coaching you can hand to the rep as-is. Change the four items to match your own methodology — the structure holds for MEDDIC stages, talk tracks, or discovery depth. The Leadership, Coaching & People Management category carries rubric-based coaching prompts by scenario.
Pipeline review prompts: interrogate the deal list before Monday
"Here is my team's pipeline export: [paste rows]. Flag every deal where the close date is inside 30 days but the last activity is older than 14 days, every deal that has slipped its close date more than once, and every deal above [amount] with a single contact. For each flagged deal, write the one question I should ask the rep in pipeline review. Output as a list ordered by risk."
This turns pipeline review from a status readout into an inspection. The model finds the inconsistencies; you spend the meeting on the six deals that need pressure instead of touring all sixty. Pipeline inspection prompts live in the Sales Operations, CRM & Productivity category.
Forecast prompts: the narrative your VP reads
"Here is my commit, best case, and pipeline data: [paste]. Write a forecast narrative of no more than 200 words: the number, what changed since last week, the two deals that decide whether we hit it, the risk I am watching, and what I need from leadership. No hedging language and no adjectives — every claim tied to a deal or a number."
A forecast narrative fails in two directions: a bare number with no story, or a story with no accountability. The word cap and the "tied to a deal or a number" constraint hold the middle.
1:1 prep prompts: from rep metrics to a focused agenda
"Here are this rep's numbers for the last 30 days and my notes from our previous 1:1: [paste]. Draft a 1:1 agenda with three items: one thing to recognize backed by a metric, one pattern to dig into with the data that shows it, and one commitment to review from last time. Frame the dig-in item as a question, not a diagnosis."
The last constraint matters most. An agenda that arrives as a diagnosis puts the rep on defense; a question invites the rep to bring the explanation you do not have.
Role-play prompts: your hardest buyer, on demand
Rep training works better when the practice pressure is real. "Play a [title] at a [industry] company evaluating our product. You are skeptical, budget-constrained, and have been burned by a similar tool. Respond to the rep's pitch one message at a time. Escalate one objection per turn, starting with [most common objection]. Do not break character until I say stop, then score the rep on how well they held the frame." Run it live in team meetings or assign it as solo practice; pair it with the prompts in ChatGPT prompts for objection handling for the response frameworks themselves.
Team communication prompts
Kickoff notes, QBR summaries, and win announcements are recurring writing tasks with recurring structures, which makes them one-paste jobs: give the model the raw material and the format you always use, and stop writing them from a blank page. The test for all of these is the same as for rep-level prompts — the output should be ready after a trim, not a rewrite. If your team is standardizing on a shared set, the comparison of prompt libraries for sales teams covers the options; browsing Promptifi is free, and paid plans are flat at $12/month, $89/year, or $149 lifetime. Start with a free prompt or see pricing.
Frequently asked questions
How can sales managers use ChatGPT for coaching?
Paste a call transcript and a scoring rubric — four or five observable behaviors — and ask the model to score each item with a supporting quote from the transcript plus one behavior to change. The rubric turns vague encouragement into evidence-based coaching, and the transcript quotes make the feedback concrete enough for the rep to act on.
Can ChatGPT help with sales forecasting?
It cannot predict which deals close, but it is effective at building the forecast narrative: paste your commit and pipeline data and ask for a 200-word summary covering the number, what changed, the deciding deals, and the risk. The discipline of tying every claim to a deal or a number is what makes the narrative credible upward.
What are good AI prompts for pipeline reviews?
Inspection prompts. Paste the pipeline export and ask the model to flag stale activity on near-term deals, repeated close-date slips, and large single-threaded deals, with one question to ask the rep per flagged deal. The model does the scan; the meeting time goes to the deals that need it.
How do I use AI to prep for 1:1s with reps?
Give the model the rep's recent metrics and your notes from the last 1:1, and ask for a three-item agenda: recognition backed by a metric, one pattern to explore framed as a question, and last session's commitment to review. Ten minutes of prep per rep becomes two, and the agenda stays about the rep rather than the number.