Using AI as a Sales Leader
Where AI actually earns its keep in a sales leader's real week, three prompts built for pipeline calls and rep coaching, and which documented sales frameworks are worth loading alongside it.
A sales leader's week runs on the tension between what the pipeline says and what's actually happening in it: a deal marked late-stage that hasn't actually moved in three weeks, a rep whose numbers are down for a reason nobody's named yet, a forecast conversation where two people are looking at the same data and reaching different conclusions. Generic AI use treats the pipeline as a spreadsheet to summarize. The more useful use treats it as a set of specific claims to check against the actual evidence behind each one.
What does a sales leader's actual week look like?
The recurring calls repeat across nearly every pipeline review: whether a deal marked late-stage actually shows the evidence late-stage deals require, multiple stakeholders engaged, a defined next step, rather than just sitting in that column because nobody moved it back. Whether a specific rep's slipping numbers trace to a skill gap, a territory problem, or a motivation issue that looks identical from the outside. Whether a forecast disagreement between two people looking at the same data is actually a data problem or a definition problem, different people quietly using different bars for what counts as committed. None of these resolve by staring at the dashboard longer, since the dashboard shows the stated status, not the underlying evidence that would confirm or contradict it.
Sales leadership carries a specific kind of pressure the individual contributor role doesn't: the leader is accountable for a forecast built from other people's read of their own deals, filtered through whatever confidence or optimism each rep happens to bring to a pipeline review. Catching the gap between a stated stage and the actual evidence behind it is a recurring, structural problem, not a one-off judgment call.
Where does generic AI actually help here, and where does it quietly fail a sales leader?
Unaided, it's genuinely useful for structuring an evidence check against a deal's stated stage once you describe what's actually happened in it, and for generating specific, evidence-grounded coaching language once you describe a rep's actual pattern rather than a vague impression of it.
It's weak at anything requiring direct observation: the actual tone of a call, what a rep chose not to write in their notes, what a prospect's internal politics really look like beyond the secondhand report. Feed it a vague summary and it will still produce a confident-sounding read.
What's worth checking with AI before this week's pipeline review?
These run in Claude, ChatGPT, or Gemini as written.
"This deal is marked as [stage]. Here's what's actually happened in it: [describe stakeholders engaged, last interaction, timeline]. Does the evidence actually support this stage, or is it more likely one stage behind?"
Why it works: checking a stated stage against the specific evidence that stage is supposed to require catches deals that are sitting in a column out of habit rather than because they actually moved there.
"Here's what I've observed in [rep]'s recent calls and deal notes: [describe it]. Based on documented research on what predicts success in complex sales, where does their pattern actually break down?"
Why it works: grounding the diagnosis in documented research on what actually predicts success, rather than a generic sense that the rep "needs to improve," produces a coaching point specific enough to actually act on.
"[Person A] and [Person B] disagree on whether this deal should be forecast as committed. Here's the data both are looking at: [describe it]. Is this a disagreement about the facts, or about where each person draws the line for "committed"?"
Why it works: most forecast disagreements aren't actually about the data, they're about an unstated difference in where each person's personal bar sits, and naming that explicitly turns an unresolvable argument into a specific, fixable definition problem.
Where do these three prompts stop helping?
Run all three and this week's pipeline review and coaching conversation will be sharper than a dashboard scan alone produces. What you won't have is a system that knows the actual tone of a call, what a rep left out of their notes, or what a prospect's internal politics genuinely look like. The gap isn't this week's forecast; it's that no prompt substitutes for the read a sales leader builds by actually listening to calls and watching a rep work a room.
Which documented frameworks actually fit a sales leader, and why?
Neil Rackham's SPIN research, built from analysis of thousands of real recorded sales calls, documented that question sequence, not feature pitching, predicts success in large complex sales, directly relevant to diagnosing why a rep's calls aren't converting. Jill Konrath's documented case for meeting an overloaded buyer with a reduced decision rather than more information reframes what "good rep behavior" actually looks like with a modern, time-poor buyer. Matthew Dixon's Challenger Sale research, conducted jointly with Brent Adamson, found that the rep who teaches the customer something new about their own problem outsells the rep who simply builds relationship, a specific, checkable behavior to coach toward.
Each targets a different recurring failure: Rackham for a rep who's pitching instead of asking, Konrath for a rep who's adding information instead of reducing the buyer's decision, Dixon for a rep who's building rapport without ever teaching the buyer anything new.
What does it take to load one of these into Claude, ChatGPT, or Gemini?
In Claude, each installs as a Skill: go to Settings → Skills → Add skill and upload the .zip. Claude reads the skill's description and auto-invokes it when your question matches, or you can force it with /neil-rackham-framework at the start of a message. Claude Code users get the same thing via ~/.claude/skills (every project) or a project's .claude/skills folder (that project only).
In ChatGPT, the plain .md file goes into a Custom GPT: Create a GPT → Configure → Instructions, paste the framework's content in. Standard Custom Instructions (the "About me" fields) cap out at 1,500 characters, too short for a full framework, so use a Custom GPT specifically, not Custom Instructions.
In Gemini, paste the .md content directly into the system prompt if the interface you're using supports one, or lead your first message with it: "Here is a thinking framework to apply throughout this conversation. Read it carefully, then answer using this framework's approach," followed by the file's contents.
Where should a sales leader actually start?
The Sales Leader category collects Rackham, Konrath, and Dixon alongside Aaron Ross and Mark Roberge, each a documented method for a different part of running a sales team. For the rep-level conversation specifically, Discovery Call Prep is a $79 tool built to structure the questions and traps before a specific call, useful to hand directly to a rep you're coaching.
All the copy-paste prompts from this guide, and the rest of the AI Practice series, live at /prompts too, free to copy without reading the guide first.
Frequently asked questions
Can AI actually tell me which deals in the pipeline are going to close?
It can help you structure a read on a deal once you describe what's actually happened in it, how many stakeholders are engaged, how the last few interactions went, whether the timeline has slipped. It has no independent visibility into the deal itself, so the read is only as reliable as your own honest account of it, not a signal it discovered on its own.
What's the single biggest mistake sales leaders make asking AI for pipeline help?
Asking it to forecast the number instead of asking it to check the evidence behind each deal's stated stage. A forecast question invites a confident-sounding guess; asking whether a deal marked as late-stage actually has the specific signals late-stage deals require, multiple stakeholders engaged, a defined timeline, surfaces the gap between the stated stage and the real one before the number gets reported upward.
Can this help with coaching a specific rep, not just the pipeline overall?
Yes, and it's one of the more directly useful applications here. Describing a specific rep's actual call patterns or deal notes and asking where their documented method is breaking down, not closing enough, not asking enough discovery questions, produces a genuinely specific coaching point, sharper and more actionable than a generic "do more discovery" note ever manages to be on its own.
Is it fair to use AI to write performance reviews for reps?
For structuring an honest, specific review from real observations, yes, and it often produces something clearer and less generic than a rushed one written from memory the night before. For inventing the underlying assessment of a rep's performance, no, since that has to come from what you've actually observed in their calls and their numbers, not from a template quietly filling in the gaps.
How is this different from the sales-call-prep guide already on the site?
That guide is built for one rep prepping for one specific discovery call, the situation map, the implication ladder, the objection library, all aimed at that single conversation. This one is built for the sales leader's broader job, reading the pipeline honestly, coaching a specific rep, and resolving a forecast disagreement, using the same underlying research from Rackham, Konrath, and Dixon.
Do these prompts work for a small team, or only enterprise sales orgs?
The underlying logic transfers regardless of team size, checking evidence against stage, diagnosing a specific rep's pattern, structuring a forecast disagreement. A two-person sales team benefits from the same discipline a fifty-person org does; there's simply less pipeline to apply it to, not a fundamentally different method required for applying it well at that smaller, earlier scale of operation the team is running at.
What's the honest limit of using AI for sales leadership decisions?
It has no access to the actual tone of a call, what a rep didn't say in their notes, or what a prospect's internal politics really look like beyond what's been reported. Use it to structure the evidence and catch a gap between a deal's stated stage and its real one; the judgment about a specific rep or a specific deal still has to be yours.
Written by Gareth Hoyle. Last updated 26 September 2026. Part of the authority.md guides library.
More guides.
Bringing AI Into a Creative Project Without Ruining It
What generic AI brainstorming actually gives a creative project and what it quietly flattens, four prompts that help today without taking over, and what changes when the AI applies Rick Rubin's documented listening practice instead of generic idea generation.
AI for Hiring Decisions: Where Claude and ChatGPT Actually Help
A grounded look at what generic AI use gets right and wrong when you're deciding on a hire, four prompts that genuinely sharpen the decision, and what changes when the AI applies a documented evidence-based hiring framework instead of generic interview advice.
Using AI to Prep for a Negotiation: What Actually Works
Generic AI negotiation advice is genuinely useful and genuinely limited. Here's what Claude and ChatGPT get right before a negotiation, four prompts that work today, and what changes once the AI has a documented framework instead of a generic one.