These are starting points, not finished tools. They work best when you:
- Modify them for your specific industry and customer profile
- Load your CRM data, win/loss history, and past proposals into a Claude Project
- Build muscle memory using the same five or six prompts every week
Don’t expect them to be perfect on first run. Iterate.
Pre-call preparation
What it’s for: A one-page briefing before a first call so you walk in informed instead of generic.
I have a call tomorrow with [company]. Help me prep.
What I know about them:
[paste anything you have — website notes, prior interactions, LinkedIn, public news]
What I want to accomplish on the call:
[paste your objective]
Draft a 1-page briefing including:
- The company in 2 sentences
- 3 specific things I should know about their business right now
- 2 likely pain points based on their profile
- 4 questions worth asking on the call
- 1 risk to be aware of going in
Be specific. Don't tell me generic things about their industry.
What it’s for: Understanding the person you’re about to meet so you can connect, not just pitch.
I'm going to be on a call with [name], [title] at [company]. Help me understand them.
Sources I have:
[paste LinkedIn, recent posts, public talks, any prior interactions]
Output:
- Their background in 2 sentences
- 3 themes that appear in their public content
- What likely matters to them in their current role
- 2 specific things to say that would resonate based on their interests
- 1 thing to avoid
Don't make things up. If you don't have evidence, say so.
Discovery and qualification
What it’s for: Generating sharper follow-up questions mid-discovery instead of falling back on a generic script.
I'm in discovery with a prospect. Here's what I know so far:
[paste discovery notes to date]
Help me identify the 5 best follow-up questions to ask. Each question should:
- Surface information I don't yet have
- Test a hypothesis I'm forming
- Be open-ended (not yes/no)
- Be specific to their business (not generic)
For each question, tell me what I'd learn if they answered it well and what to listen for in the answer.
What it’s for: Scoring a prospect honestly against your qualification criteria so you spend time on the right deals.
Score this prospect on our qualification criteria. Be honest — don't give points the data doesn't support.
Prospect data:
[paste discovery notes, public info, financial data if available]
Score on (1-5 each):
- Fit (do they match our ICP?)
- Pain (do they have a real, named problem?)
- Authority (am I talking to the decision maker or close to them?)
- Budget (can they afford this?)
- Timing (is this happening this quarter / year / never?)
- Sponsor (is there an internal champion?)
Total /30. Above 22 = pursue actively. 15-22 = nurture. Below 15 = disqualify or wait.
Justify each score with 1 specific data point.
Proposal and quote drafting
What it’s for: Turning raw discovery notes into a skimmable first-draft proposal that mirrors back what the buyer told you.
Draft a proposal based on the following discovery notes. Use our standard proposal structure:
- Their situation as we understand it
- The work we're recommending
- The expected outcomes
- The investment
- The timeline
- Why us
- Next steps
Discovery notes:
[paste discovery notes]
Style: direct, specific, matches what they said back to them so they feel heard. Avoid jargon. Avoid hedging language. Avoid "best-in-class" or similar consultant language.
Length: ~3 pages laid out. Bullet-heavy. Easy to skim.
What it’s for: Building a pricing rationale that survives internal scrutiny and gives the buyer language to defend it.
I'm pricing this engagement at [price]. Help me articulate the rationale in a way that survives scrutiny.
Engagement details:
[paste scope, team, timeline]
Comparable engagements we've priced this way:
[paste 2-3 prior similar engagements with their pricing if available]
Output:
- 1-paragraph rationale (why this price for this scope)
- 3 specific points justifying the price
- 1 paragraph addressing the most likely objection
- An ROI frame the buyer can use to defend it internally
Don't apologize for the price. Justify it.
Pipeline review
What it’s for: A reality check on a single deal’s forecast before it goes into the commit.
Review the following deal and tell me whether the rep's forecast is realistic.
Deal data:
[paste opportunity details, stage, close date, recent activity, stakeholder map]
Output:
- The most likely close date (your best read)
- The most likely outcome (won / lost / pushed)
- Confidence level (high / medium / low)
- 2 specific risks
- 2 things that would materially change the forecast in the next 30 days
Be honest. Don't be optimistic to please.
What it’s for: Surfacing patterns across the whole pipeline that you’d miss deal-by-deal.
Analyze our pipeline and surface patterns I might be missing.
Pipeline data:
[paste pipeline export — deals, stages, age, size, source, owner]
Output:
- 3 patterns I should know about (could be positive or concerning)
- 1 stage where deals are stalling more than expected (with evidence)
- 1 segment where we're winning above our average (with evidence)
- 1 specific recommendation for action this week
Don't fabricate. If the data doesn't support a pattern, say so.
Customer expansion
What it’s for: Finding specific, evidence-backed expansion plays inside an existing account.
This is a current customer. Identify expansion opportunities.
Account data:
[paste customer info, current contract, usage, recent interactions, stated priorities]
Output:
- 3 specific expansion opportunities (not generic ones — specific to what they've told us or what their data suggests)
- For each: the case, the rough size of opportunity, the right internal champion, the right entry point
- The single best one to pursue this quarter, with reasoning
What it’s for: Framing a renewal conversation 90 days out, grounded in the value actually delivered.
This customer has a renewal coming up in 90 days. Help me position the renewal conversation.
Customer data:
[paste account info, value delivered, recent friction, expansion potential]
Output:
- Renewal recommendation (renew as-is, upgrade, downgrade, walk)
- The narrative for the renewal conversation (3-4 sentences)
- 2 specific moments of value to call out from the past year
- 1 thing we'd commit to differently going forward
- The right ask if upgrading is appropriate
Lost deal analysis
What it’s for: An honest post-mortem on a loss so the lesson sticks and the deal isn’t written off prematurely.
We lost this deal. Help me understand why honestly.
Deal data:
[paste deal history, stage progression, stakeholder interactions, lost reason given]
What the customer said:
[paste any verbatim from the customer about why]
Output:
- The stated reason for the loss
- 2 likely underlying reasons (your read, even if uncomfortable)
- 1 thing we'd do differently if we ran the deal again
- 1 thing we did well that should be preserved
- A read on whether the deal is genuinely lost or could be reactivated in 6-12 months
What it’s for: Drafting a reactivation email for a deal that went dark — without pretending it didn’t.
This deal went dark 4 months ago. Draft a reactivation email.
Deal context:
[paste deal history, last contact, the reason we think it went dark]
Email requirements:
- 100-150 words
- Doesn't apologize for being out of touch
- References something specific from the prior conversation
- Brings new value (a relevant case study, a market observation, an insight)
- Has a soft ask (a 20-minute conversation, not "ready to move forward?")
- Plain language, no "circling back" or "touching base"
Goal: reignite the conversation without pretending the last 4 months didn't happen.
A note on what these prompts don’t do
These prompts make a competent salesperson more leveraged. They don’t make an incompetent salesperson competent. The judgment, the relationship, the trust — those are still human work.
They also don’t replace your CRM, your sales process, or your forecasting discipline. They speed up the cognitive work inside those systems.
Want a structured frame for this?
If your sales team is starting to use AI seriously and you’d like a structured frame for what to build — and what not to — that’s a useful 30-minute call. It’s the kind of focused, scoped work an AI Office retainer is built around.