March 21, 2026

Close More Deals: AI Prompts for Sales Teams

Here's a number that should bother you: the average sales rep spends less than a third of their day actually selling. The rest is emails, research, updating the CRM, writing follow-ups, prepping for calls they already know how to run. Admin work that feels productive but isn't quota.

AI is genuinely good at most of that admin. Not "good enough" — actually good. But the prompts you use matter enormously. A vague prompt gets you a generic result you'll rewrite anyway. A specific prompt with real context gets you something you can send in 30 seconds. This post walks through every stage of the sales process with prompts that work.

One rule before we start: always include prospect context in your prompts. Company name, industry, size, what they do, what pain you think they have. The more specific you are going in, the less editing you do coming out.

Prospecting

Cold outreach lives and dies on relevance. Generic emails get ignored. The goal here is to give AI enough context about the prospect that it can write something that doesn't read like a template — even if you sent a thousand of them.

Write a cold email to [first name] at [company]. They're a [title] at a [industry] company with about [headcount] employees. Their company recently [trigger event — e.g., raised a Series B / expanded into a new market / posted a job for X role]. We sell [product] which helps [persona] with [problem]. Keep it under 100 words, no buzzwords, one clear ask.
I'm prospecting into [company]. Based on what I know — [2-3 bullet points about the company] — what business problems might they be facing that [our product] could help with? Give me 3 angles I could use in outreach.

Discovery

Discovery calls go sideways when you run out of questions or let the prospect ramble without direction. AI can help you prepare a tight question set that actually uncovers pain — not just a list of generic "what keeps you up at night" prompts.

I have a discovery call with [name], [title] at [company]. They came inbound through [channel] and their initial interest was in [topic]. We sell [product]. Write me 8 open-ended discovery questions focused on uncovering [specific pain area], their current process, and decision-making structure. Skip anything generic.

After the call, flip it around and use AI to summarize what you heard and identify gaps:

Here are my notes from a discovery call: [paste raw notes]. Summarize the prospect's key pain points, their current solution, what's driving urgency (if anything), and what's still unclear. Flag any red flags.

Pitch & Demo

Nobody wants to sit through a demo that isn't about them. Before every pitch, spend five minutes prompting an AI to help you customize your talk track to the specific company and persona.

I'm demoing [product] to [title] at [company] in the [industry] space. Their main pain points (from discovery) are: [list]. Write a 3-minute talk track that leads with their problem, shows how we solve it specifically, and ends with a trial close. Avoid feature-dumping.

Objection Handling

Most objections are predictable. "It's too expensive." "We're happy with our current vendor." "Now isn't the right time." You've heard them all. AI is great for preparing calm, confident responses before you're in the moment.

A prospect told me "[exact objection]." We sell [product] at [price point]. Their context: [1-2 sentences about their situation]. Write 3 different responses I could give — one that acknowledges and reframes, one that asks a question to dig deeper, and one that offers a path forward. Keep each under 4 sentences.

Run this for every objection you hear repeatedly. Build a personal library of responses. You'll start internalizing the frameworks without needing to prompt at all.

Follow-up

The fortune is in the follow-up — and also where deals go to die because reps either give up too early or send the same "just checking in" email five times. Neither works.

Write a follow-up email after a demo with [name] at [company]. The demo went well but they said they needed to [next step — e.g., loop in their CFO / review with the team]. Key things they responded to: [list]. I want to reference something specific we discussed, provide one piece of value (a relevant case study, stat, or resource), and propose a concrete next step. Keep it short.

For deals that have gone quiet, give AI the full context of the thread and ask it to write a breakup email that re-opens the door without being desperate. It's surprisingly good at those.

Reporting & Pipeline Reviews

Reporting is where sales time goes to die. Pipeline reviews, forecast calls, deal summaries for your manager — all of it eats into selling time. AI can cut this dramatically.

Here are my open opportunities: [paste deal names, stages, values, and any notes]. Write a concise pipeline summary for my weekly review. Flag which deals need attention this week, which are at risk, and what actions I should prioritize.

The one thing that makes all of this work

Personalization. Every single category above gets dramatically better when you give the AI real context. Don't just say "prospect." Say "VP of Operations at a 200-person logistics company in the Midwest who told me their biggest problem is driver retention." The AI can only work with what you give it.

A good habit: keep a running note for each active deal with key facts about the prospect. Company info, what they care about, what they said on the last call. Paste that into every prompt. You'll get outputs that are half-finished work instead of first drafts.

If you want a place to write your prompt once and try it across ChatGPT, Claude, and Gemini at the same time — different models handle tones differently and it's worth comparing — that's exactly what Prompt Router is for. And the Sales Prompts library has a full set of ready-to-use prompts you can grab and customize in seconds.

Ready-to-use prompts for every stage of the sale.

Browse Sales Prompts    Open Prompt Router