92% of sales reps report using AI tools (only about 8% use none), and 83% say AI helps them personalize prospect interactions, per HubSpot's State of Sales research.Source ↗
AI adoption among sales reps rose from 24% in 2023 to 43% in 2024 — a roughly 79% year-over-year jump.Source ↗
Sellers who effectively partner with AI tools are reported to be 3.7x more likely to meet quota than those who don't.Source ↗
Reps spend only about 25% of their time actually selling; logging activity, email, and data entry consume most of the rest — the admin AI is best suited to reduce.Source ↗
writingClaudeChatGPT

Personalizing first-touch outreach from real, cited buying signals

The default move — asking AI for a "personalized" cold email — produces the exact opener every buyer now recognizes and deletes: the vague compliment, the "love what you're doing in the space" line, the template that could go to anyone. With 92% of reps using AI, inboxes are saturated with that pattern and reply rates have fallen. What still works is grounding a short message in one specific, verifiable trigger you supply.

Prompt
You are helping a B2B sales rep write a first-touch outreach email. Write ONE short email to a prospect based only on the real signal I give you — not generic flattery.

Prospect role and company: {{prospect_role}}
The specific, verifiable signal I'm reaching out about: {{signal}}
What I sell and the one problem it solves for them: {{my_offer}}
The single call to action: {{cta}}

Requirements:
- Under 90 words. One clear ask. Conversational, peer-to-peer, no hype.
- Open by referencing the specific signal above and why it makes my solution relevant right now — do NOT open with a generic compliment.
- Use only facts I gave you. If a detail would strengthen the email but I didn't provide it, insert "[VERIFY: what to confirm]" instead of inventing it. Never invent a stat, a mutual connection, or a customer name.
- End with a plain one-line opt-out (e.g., "Not the right time? Just reply 'no' and I'll stop reaching out.") and leave "[COMPANY MAILING ADDRESS]" on its own line for CAN-SPAM compliance.
- Give me two subject-line options, each honest and specific to the signal — no clickbait.

Fill in your details and the prompt updates live — then copy.

analysisClaudeChatGPTGemini

Turning a prospect's public filings and news into buying signals

Good outreach and good discovery both depend on knowing what's actually happening inside the account — its priorities, pressures, and initiatives — but reading an earnings call transcript, a press release, and a stack of job posts can eat an hour before you've written a word. AI is strong at compressing that public material into signals and talking points, as long as you feed it only public text and verify every quote before repeating it.

Prompt
You are a B2B sales research analyst. Below is public material about a prospect company. Extract sales-relevant signals from it — nothing more.

Company: {{prospect_company}}
What I sell / the problem area I help with: {{my_solution_area}}
Public material (earnings call excerpt, press release, 10-K section, job posts, blog):
{{public_materials}}

Produce:
1. Top 3 business priorities you can see in the text, each with the exact quoted line that supports it.
2. Likely pains in my solution area that connect to those priorities — labeled as hypotheses to test, not facts.
3. Three specific talking points I could open a conversation with.
4. Three discovery questions tailored to what you found.

Hard rules:
- Use ONLY the pasted text. If a priority isn't supported by a direct quote, don't list it.
- Never infer revenue, headcount, budget, or private plans that aren't stated. Where something isn't in the text, write "not stated."
- Do not paraphrase a quote into something the source didn't say. Mark anything I should double-check with "[VERIFY]".

Fill in your details and the prompt updates live — then copy.

planningClaudeChatGPT

Building a discovery-call prep sheet and question plan

Solid discovery prep takes 30-60 minutes per call, and reps running 6-8 meetings a day simply can't do it, so they wing it and ask the same generic questions. AI can build the agenda, a layered set of discovery questions tied to the account's likely priorities, and the objections you should expect — turning what you already know into a one-page plan you can actually run the call from.

Prompt
You are an expert sales coach helping me prep a discovery call. Build a one-page prep sheet.

Account context (de-identified — no names): {{account_context}}
Deal stage and what I already know: {{deal_stage}}
What I most need to learn on this call: {{goal_of_call}}

Produce:
1. A five-point agenda for a 30-minute call.
2. 8-10 discovery questions, layered from broad to specific, mapped to a qualification framework (BANT or MEDDIC). Favor open questions that surface pain, priorities, decision process, and timeline.
3. Three objections I'm likely to hear, each with a one-line reframe.
4. Three things to confirm before I can advance the deal.

Rules:
- Phrase everything about their business as a question to ask, never as a fact you're asserting — you have not verified their situation.
- Do not invent company news, competitors, or budget figures. If prep would benefit from research I haven't done, mark it "[RESEARCH]".
- Keep it scannable: short lines, no paragraphs.

Fill in your details and the prompt updates live — then copy.

communicationClaudeChatGPT

Drafting objection-handling follow-ups that move a stalled deal

Most sales take five or more touches, and the follow-up after an objection — too expensive, bad timing, need to check with the team — is where deals stall because reps freeze on the wording. A strong follow-up acknowledges the objection, reframes it, offers a relevant proof point, and proposes a specific next step. AI drafts that fast, but every proof point and price has to come from you.

Prompt
You are helping a B2B sales rep write a follow-up email after a prospect raised an objection. Draft two short variations.

The objection, in the prospect's words: {{objection}}
Deal context (de-identified): {{deal_context}}
Approved proof points I may use — real case studies, metrics, or references I'm giving you: {{approved_proof}}
The next step I want: {{desired_next_step}}

For each version:
- Under 120 words. Acknowledge the objection genuinely (no dismissing it), reframe it, offer one relevant proof point, propose the next step.
- Warm and consultative. No pressure tactics, false urgency, or guilt.

Hard rules:
- Use ONLY the proof points, customer names, and numbers I provided above. Never invent a case study, a reference customer, an ROI figure, or a statistic.
- Do not promise pricing, discounts, timelines, or contract terms I did not give you. If a claim would help but I didn't provide it, write "[NEED: what to confirm]".
- Include a one-line opt-out, since this is commercial email.

Fill in your details and the prompt updates live — then copy.

automationClaudeChatGPTCopilot

Turning messy call notes into clean CRM records and next steps

Reps spend the majority of their time on non-selling work, and for someone running 40 calls a day, after-call admin alone can hit two hours. The result is fragmented notes that never become clean CRM entries, so pipeline data and forecasts rot. AI turns raw, half-typed notes into a structured summary, populated qualification fields, and dated action items in seconds.

Prompt
You are a CRM documentation assistant for a sales rep. Turn my rough call notes into a clean CRM record. Accurate pipeline data matters more than polish.

Call type: {{call_type}}
My raw notes (customer identifiers already removed): {{raw_notes}}

Produce:
1. SUMMARY — Situation / Pain / Decision process / Timeline / Next steps, in past tense and neutral tone.
2. QUALIFICATION — fill BANT (Budget, Authority, Need, Timeline); write "not discussed" for anything I didn't mention.
3. ACTION ITEMS — numbered, each with an owner and a due date drawn from my notes.
4. OPEN QUESTIONS — anything ambiguous.

Hard rules:
- Include ONLY what is in my notes. Never invent a budget, a stakeholder, a commitment, or a timeline I did not write down.
- If a fragment is ambiguous, put it under OPEN QUESTIONS marked "[CONFIRM]" rather than guessing what I meant.
- Do not soften or upgrade the deal's status — if I wrote that they're hesitant, keep it hesitant.

Fill in your details and the prompt updates live — then copy.

writingClaudeChatGPT

Drafting a first-pass proposal from your discovery notes

Writing a proposal from a blank page after discovery takes hours, so reps put it off and the deal cools. AI can assemble a structured first draft that mirrors the prospect's own language and your approved boilerplate — but scope, pricing, SLAs, and any ROI claim have to be yours, because a proposal is close to a contract and the AI will happily invent numbers.

Prompt
You are helping a B2B sales rep draft a first-pass proposal. Build a structured draft from my inputs only.

Prospect's situation and goals (from discovery): {{prospect_situation}}
Approved solution details, scope, and pricing I'm giving you (use exactly as written): {{approved_solution_and_pricing}}
Timeline: {{timeline}}
What makes us different (approved): {{differentiators}}

Produce a proposal with these sections: Understanding of your situation; Your goals; Proposed solution; Scope of work; Timeline; Investment; Why us; Next steps.

Hard rules:
- Use ONLY the capabilities, scope, pricing, SLAs, and terms I provided. Never invent a feature, a price, a guarantee, a discount, or an ROI figure.
- Leave the Investment section's pricing exactly as I wrote it — do not reformat, round, or "estimate" any number.
- Where I haven't given you a needed detail, insert "[TBD]" rather than filling the gap.
- Mirror the prospect's own words from the discovery notes. Plain, confident language — no hype words like "transform" or "revolutionize".

Fill in your details and the prompt updates live — then copy.

Common questions from sales reps

Is it legal to send AI-written cold emails to businesses I've never contacted?

In the U.S., yes — CAN-SPAM doesn't require prior consent for B2B outreach, so you can email a business contact cold. But every commercial message still needs honest From and subject lines, a valid physical postal address, and a working opt-out that you honor within 10 business days, whether a human or AI wrote it. Other countries are stricter — the EU's GDPR and Canada's CASL generally require consent — so check the recipient's jurisdiction.

Can I put customer or deal information into ChatGPT?

Not into consumer versions. Customer names, contact details, contract terms, pricing, and pipeline data are confidential, and free or personal-tier tools may retain what you paste and use it to train the model. De-identify your notes first, or use an enterprise or zero-retention tool your company has approved with a proper data agreement.

Why are my AI-personalized emails getting fewer replies, not more?

Because the easy AI output is generic — the templated compliment and "love what you're doing" opener that every buyer has now seen thousands of times and deletes on sight. Personalization only works when it's anchored to a specific, verifiable signal the prospect knows a competitor couldn't have guessed. Feed the AI one real trigger; don't let it write something it could send to anyone.

Will AI replace sales reps and SDRs?

The evidence points to leverage, not replacement. Reps who partner well with AI are reported to be far more likely to hit quota, largely because AI absorbs the research, note-taking, and admin that consume most of a seller's week. The parts that close deals — building trust, reading a room, negotiating, and judgment about what to say and what never to automate — stay human. AI drafts and organizes faster; it doesn't own the relationship.

Related professions