AI Prompts for Sales: 10 Templates for Outreach, Follow-ups, and Closing

Sales reps spend hours writing emails that get ignored, prepping for calls they wing anyway, and updating CRMs with notes that say "good conversation." AI can handle all of that — if you give it the right prompts.

The difference between a sales rep who gets value from AI and one who doesn't isn't the tool. It's the prompt. Vague inputs like "write a cold email" produce generic output that sounds like every other AI-generated email in your prospect's inbox. Structured prompts produce messages that actually get replies.

Here are 10 templates for the tasks sales teams do every day — tested across GPT-4o, Claude, and Gemini.

1. Cold Outreach Email (Personalized from Research)

The biggest mistake in AI-generated cold emails: they sound like AI-generated cold emails. This template forces personalization by requiring specific prospect details.

Write a cold outreach email.

Prospect: [Name], [Title] at [Company]
What they do: [one sentence about the company]
Something specific I noticed: [recent news, LinkedIn post,
company announcement, or mutual connection]
What we sell: [your product — one sentence]
How it's relevant to them: [specific pain point or use case]

Rules:
- Under 100 words (short emails get 2x reply rates)
- First line references the specific thing I noticed — NOT
  "I came across your company" or "I hope this finds you well"
- One clear value proposition in the middle
- CTA: low-commitment question, not "book a demo"
- No flattery ("huge fan of your work")
- No buzzwords ("leverage," "synergy," "revolutionary")
- Sign off with first name only

Why it works: The "something specific I noticed" field is the key. It forces you to do 30 seconds of research before generating — and that research is what separates emails that get replies from emails that get deleted.

2. Follow-Up Sequence (After No Reply)

One cold email isn't enough. Most deals close after 5-8 touchpoints. But each follow-up needs a different angle — not just "checking in."

Write a 4-email follow-up sequence for a prospect who didn't
reply to my initial cold email.

Original email context: [paste your first email or summarize it]
Prospect: [Name], [Title] at [Company]
Our product: [what you sell]
Their likely pain point: [what problem you solve for them]

Requirements for each email:
- Different angle (don't repeat the same pitch)
- Under 75 words each
- Increasing directness with each email
- No guilt-tripping ("I haven't heard back")
- No "just checking in" or "circling back"

Sequence structure:
- Email 1 (Day 3): Add new value — share a relevant insight,
  stat, or case study
- Email 2 (Day 7): Social proof — brief mention of a similar
  company you helped
- Email 3 (Day 14): Direct question — ask about their current
  approach to [problem]
- Email 4 (Day 21): Breakup email — acknowledge they're busy,
  leave the door open, no pressure

For each email, provide subject line + body.

Why it works: The sequence structure prevents the AI from writing four versions of the same email. Each email has a defined purpose and angle — value, proof, curiosity, respectful close.

3. Discovery Call Question Generator

Going into a discovery call with generic questions like "What are your biggest challenges?" wastes the prospect's time. This template generates questions tailored to the specific deal.

Generate 12 discovery call questions for this prospect.

Prospect: [Name], [Title] at [Company]
Industry: [their industry]
Company size: [employees / revenue if known]
What we sell: [your product]
What we know so far: [any intel from previous conversations,
their website, LinkedIn, etc.]
Deal stage: [first call / follow-up / technical evaluation]

Generate questions in three categories:

Situation (4 questions):
- Understand their current setup, tools, and processes
- Don't ask things you could find on their website

Pain (4 questions):
- Uncover specific problems and their impact
- Ask about cost of the problem (time, money, headcount)
- Get them to quantify the pain

Decision (4 questions):
- Understand their buying process and timeline
- Identify other stakeholders involved
- Surface potential blockers or competitors

Rules:
- Open-ended questions only (no yes/no)
- No jargon from OUR product — use THEIR language
- Each question should lead naturally to the next
- Include one "uncomfortable truth" question they probably
  haven't been asked before

Why it works: The three-category structure (Situation → Pain → Decision) follows proven sales methodology. The "uncomfortable truth" question is where the real insights come from — the questions competitors aren't asking.

4. Post-Demo Follow-Up Email

The email after a demo is where most deals stall. This template turns a generic "thanks for your time" into a deal-advancing message.

Write a post-demo follow-up email.

Prospect: [Name], [Title] at [Company]
Demo date: [when]
What we showed: [features/workflows demonstrated]
Their main pain point discussed: [what problem resonated most]
Their objection or concern: [what they pushed back on]
Next step we agreed on: [if any]
Other stakeholders mentioned: [names/roles if discussed]

Structure:
1. Open: Reference something specific THEY said during the demo
   (not what we showed — what they reacted to)
2. Recap: 2-3 bullet points of how our product addresses their
   specific pain point (not a feature list — connect features
   to their problem)
3. Address the objection: Briefly counter their concern with
   evidence (case study, data, or clarification)
4. Next step: Propose a specific action with a specific date
5. Optional: If other stakeholders were mentioned, offer to
   include them

Rules:
- Under 200 words
- No "as discussed" or "per our conversation"
- Don't resell the product — they already saw it
- Be specific, not generic

Why it works: "Reference something specific THEY said" is the key line. It proves you listened, not just presented. And addressing the objection proactively (instead of waiting for it to become a blocker) keeps the deal moving.

5. Objection Handling Script

When a prospect says "it's too expensive" or "we already have a solution," most reps wing it. This template prepares you for the top objections before they come up.

I sell [product] to [target audience].

Generate handling scripts for these common objections:

1. "It's too expensive / over our budget"
2. "We already use [competitor]"
3. "I need to talk to my [boss/team/IT]"
4. "We're not looking at this right now"
5. "Can you send me some materials and I'll review?"

For each objection, provide:

**Acknowledge:** One sentence that validates their concern
without agreeing (don't say "I understand" — say something
that shows you've heard this before and it's reasonable)

**Reframe:** A question that shifts the conversation from
the objection to the underlying problem

**Evidence:** One proof point (customer stat, case study
reference, or industry data)

**Bridge:** A sentence that moves toward the next step
without being pushy

Rules:
- Conversational tone — these are spoken, not written
- Never argue or dismiss the objection
- Each script should be under 60 words total
- No "that's a great question" or "I'm glad you asked"

Why it works: The four-step structure (Acknowledge → Reframe → Evidence → Bridge) prevents the two common failures: getting defensive or immediately discounting. This is a variation of the STOKE framework — structured prompts consistently outperform freeform instructions. The reframe question is where the real skill is — turning "too expensive" into a conversation about ROI.

6. Competitive Battle Card

When a prospect mentions a competitor, you need to respond with confidence — not scramble. This template builds a battle card you can reference in seconds.

Create a competitive battle card.

Our product: [name + one-sentence description]
Our pricing: [model and price point]
Our key differentiators: [3 things we do better]

Competitor: [name + one-sentence description]
Their pricing: [if known]
Their strengths: [be honest — what they do well]

Generate:

**Quick positioning statement** (one sentence I can say on a call):
"Unlike [competitor] which [their approach], we [our approach],
which means [specific benefit for the prospect]."

**3 trap questions** I can ask that expose their weaknesses:
- Questions where our answer is strong and theirs is weak
- Must sound natural, not like I'm attacking them

**Switching objection handler:** Script for when they say
"We'd have to migrate from [competitor] and that's too much work"

**Win story:** A 3-sentence case study of a customer who switched
from [competitor] to us (use this template: problem → switch
reason → result with numbers)

Rules:
- Never trash-talk the competitor directly
- Focus on what WE do, not what THEY don't
- Be factual, not emotional
- Acknowledge their strengths — it builds credibility

Why it works: "Never trash-talk" and "acknowledge their strengths" are counterintuitive but essential. Reps who badmouth competitors lose trust. The trap questions are subtle — they expose gaps without attacking.

7. CRM Deal Summary / Call Notes

After every call, reps should update the CRM. Most don't because it's tedious. This template generates structured notes from a quick brain dump.

I just finished a sales call. Turn my rough notes into a
structured CRM update.

My rough notes:
[paste your messy notes, bullet points, or stream of
consciousness from the call]

Format the output as:

**Call Summary** (2-3 sentences — what happened)

**Key Discovery:**
- Pain point: [what problem they have]
- Current solution: [what they use today]
- Impact: [cost/time/risk of the problem]
- Budget: [any budget info mentioned]
- Timeline: [urgency level or decision date]
- Decision maker: [who decides + who influences]

**Objections Raised:**
- [objection 1 + how I addressed it]

**Commitments Made:**
- By us: [what we promised to send/do]
- By them: [what they agreed to]

**Next Step:** [specific action + date]

**Deal Health:** [Green / Yellow / Red + one-sentence justification]

Rules:
- Only include information from my notes — don't invent details
- If something is missing (e.g., budget not discussed), write
  "Not discussed" — don't guess
- Keep it factual, not optimistic

Why it works: "Don't invent details" and "keep it factual, not optimistic" prevent the AI from turning a mediocre call into a promising-sounding CRM entry. The structured format ensures every field in your CRM gets filled accurately.

8. Proposal / SOW Draft

Writing proposals from scratch takes hours. This template generates a structured first draft you can customize in minutes.

Draft a proposal for this deal.

Client: [Company name]
Contact: [Name, Title]
Their problem: [1-2 sentences on what they need to solve]
Our solution: [which product/tier/package]
Pricing: [price point + billing model]
Timeline: [implementation / delivery timeline]
Special terms discussed: [discounts, pilot period, custom scope]

Proposal structure:

1. **Executive Summary** (3-4 sentences)
   - State their problem, our solution, expected outcome
   - No company history — they know who we are

2. **Scope of Work**
   - What's included (bullet points)
   - What's NOT included (important for preventing scope creep)

3. **Timeline & Milestones**
   - Phase 1, 2, 3 with dates
   - Key deliverables per phase

4. **Investment**
   - Pricing table
   - Payment terms
   - What happens after the contract period

5. **Next Steps**
   - What they need to do to move forward
   - Specific contact for questions

Rules:
- Professional but not corporate — clear, direct language
- Under 2 pages equivalent (~800 words)
- No filler paragraphs about "our commitment to excellence"
- Focus on THEIR outcomes, not our features

Why it works: "What's NOT included" is the most important line in any proposal — it prevents scope creep and sets clear expectations. "Focus on THEIR outcomes" prevents the common failure of proposals that read like product brochures.

9. Pipeline Prioritization Analysis

When you have 30 open deals and limited time, you need to know which ones to focus on. This template turns your pipeline data into an action plan.

Analyze my sales pipeline and tell me where to focus this week.

Here are my open deals:
[paste deal list with: company, deal size, stage, days in stage,
last activity date, next step, and any notes]

Analyze for:

1. **Deals at risk** — no activity in 14+ days, stuck in stage,
   or missing next steps. For each: why it's at risk + specific
   action to unstick it.

2. **Deals to prioritize** — highest probability of closing this
   month based on stage, engagement, and deal size. Rank top 5
   with reasoning.

3. **Deals to deprioritize** — low probability, small deal size,
   or too early stage to warrant time this week. Be honest —
   don't tell me everything looks great.

4. **Missing information** — deals where I'm missing critical
   data (decision maker, budget, timeline) that I need to uncover.

Output format: action-oriented bullets. For each deal mentioned,
include ONE specific next action I should take TODAY.

Rules:
- Be direct — if a deal is dead, say so
- Prioritize by impact (deal size × probability), not just stage
- No motivational language ("you've got a great pipeline!")

Why it works: "Be honest" and "if a deal is dead, say so" prevent the AI from producing an optimistic pipeline review. The specific "one action TODAY" per deal turns analysis into immediate action.

10. Win/Loss Analysis

Understanding why you win and lose deals is how you improve. This template turns raw deal data into actionable insights.

Analyze these closed deals and identify patterns.

Won deals:
[paste 5-10 recent wins with: company, deal size, sales cycle
length, competitor mentioned, key objection, and what sealed it]

Lost deals:
[paste 5-10 recent losses with: company, deal size, how far
they got, competitor they chose, reason they gave, and what
you think the real reason was]

Analyze:

1. **Win patterns:**
   - Common traits of companies that buy (size, industry, pain point)
   - Most effective selling points (what resonated)
   - Average winning sales cycle length
   - Competitive wins: what we did differently

2. **Loss patterns:**
   - Where in the pipeline deals die (which stage)
   - Common objections in lost deals
   - Competitor loss analysis (who we lose to and why)
   - "Real reason" vs "given reason" — any gaps?

3. **Recommendations:**
   - 3 specific changes to improve win rate
   - Which types of deals to pursue more aggressively
   - Which types of deals to qualify out faster
   - One script/messaging change based on the data

Rules:
- Be data-driven — reference specific deals to support conclusions
- Don't sugar-coat losses — honest analysis is more useful
- Recommendations must be specific and actionable, not generic
  ("improve follow-up" is useless — "send case study within
  24 hours of demo" is actionable)

Why it works: The split between "reason they gave" and "what you think the real reason was" is where the real insights live. Prospects rarely tell you the truth about why they didn't buy. The AI helps you spot the pattern across multiple deals.

Using These Templates Effectively

Customize once, reuse forever. The product description, pricing, and key differentiators stay the same across templates. Fill those in once and save the templates for daily use.

Chain templates together. Use #1 (cold outreach) → #2 (follow-up sequence) → #3 (discovery questions) → #4 (post-demo follow-up) as a deal progresses. Each output feeds the next prompt.

Add your wins. When a template produces a reply or books a meeting, save what worked. The best sales prompts evolve from your own real results — using few-shot prompting to include winning examples makes future outputs even stronger.

Don't send AI output raw. Every email, script, and proposal should get a 30-second human review. Add a personal detail, cut a sentence that feels off, make sure the facts are right. AI writes the first 90% — you add the 10% that makes it yours.

---

Want these templates automatically optimized for your specific AI model? Try Promplify free — paste any sales prompt and get a structured, framework-enhanced version that produces better output across GPT-4o, Claude, and Gemini.