AI Sales

ChatGPT for Sales: Use Cases, Prompts, and Limits

Daniel Hayes
5 min read
Cover image for a guide to using ChatGPT for sales

Key takeaways

  • ChatGPT is strong at the work around selling: pre-call research, drafting and rewriting emails, objection prep, and summarizing notes.
  • Output quality depends on context, so a prompt with no input produces the same generic, templated writing everyone else gets.
  • Its limits are hard: it invents facts, has no access to your pipeline, and does not act on anything it drafts.
  • Use it for drafting and thinking, verify what it tells you, and let a connected tool handle the execution.

ChatGPT has quietly become a daily tool for a lot of salespeople, mostly for the unglamorous work around selling: drafting the email, prepping for the call, turning messy notes into something usable. Used well, it gives a rep back an hour a day. Used badly, it produces generic output that a buyer spots in a second.

This guide covers where ChatGPT genuinely helps in sales, the prompts that get useful results, and the limits you need to respect before you lean on it.

Where ChatGPT helps in sales

The honest framing is that ChatGPT is good at the work around the conversation, not the conversation itself. These are the use cases that hold up.

Pre-call research and prep

Paste in a prospect's public bio, their company's about page, and a recent announcement, and ask ChatGPT to summarize what matters for your meeting and suggest questions to ask. It will not replace real account-based sales research, but it compresses the reading.

Drafting and rewriting emails

This is the most common use. ChatGPT is a strong first-draft writer for outreach, follow-ups, and replies, especially when you give it the context. It is weakest when you ask for a cold email with no input, because then it writes the same template everyone else gets.

Handling objections

Describe the objection you keep hearing and ask for three ways to respond, then pick and adapt. It is a useful sparring partner for a sales playbook, not a script to read verbatim.

Summarizing notes and calls

Paste rough call notes and ask for a clean summary, the next steps, and a follow-up email. Turning five lines of shorthand into a tidy recap is exactly the kind of task it does well.

Learning and roleplay

Ask ChatGPT to play a skeptical buyer and practice a discovery call, or to explain a methodology like SPICED with examples. For ramping a new rep, it is a patient, always-available coach.

Prompts that actually work

The difference between a useful answer and a generic one is almost always the prompt. A few patterns that help.

Give it a role and context.

You are an SDR selling a payroll tool to HR leaders at 200-person companies. Here is the prospect's LinkedIn summary and their company's latest news. Draft a four-sentence cold email that references the news and asks for a call.

Ask for options, not one answer.

Give me three subject lines and three opening lines for this email, ranging from direct to curiosity-driven.

Make it edit your work, not replace it.

Here is my follow-up email. Tighten it to half the length, keep the specific detail about their migration, and make the ask clearer.

Feed it the objection and the context.

A prospect said we are too expensive versus a named competitor. We are pricier but include onboarding they charge for. Give me two honest ways to respond that do not disparage the competitor.

The pattern is the same throughout: the more real input you give, the less generic the output. A prompt with no context produces a cold email that reads like every other AI sales assistant wrote it.

The limits you have to respect

ChatGPT is a drafting and thinking tool, not a system of record or a source of truth. The failure modes are predictable.

It makes things up. It will state a fact, a statistic, or a feature confidently when it has no basis for it. Never send a claim, a number, or a competitor comparison it produced without checking it yourself.

It does not know your pipeline. On its own, ChatGPT has no access to your CRM, your deals, or your past conversations, so it cannot tell you who to follow up with or score buyer intent. It writes the email; it does not know which email to send.

It is generic without input. With no context it reverts to the average of everything it has seen, which is exactly the templated tone buyers have learned to ignore.

It does not act. It drafts a follow-up, but a person still has to remember to send it, log it, and update the record. The follow-up only happens if someone acts on the draft.

Watch what you paste. Treat customer data and anything confidential with care, and follow your company's policy on what can go into a public AI tool.

ChatGPT versus a sales-specific AI tool

It helps to be clear about what ChatGPT is and is not in a sales stack.

ChatGPT is a general assistant you drive by hand: you give it context, it gives you a draft, you act. That is perfect for research, writing, and prep, and it costs almost nothing to try.

Sales-specific AI tools do a different job. A conversational AI for sales tool or an autonomous CRM is connected to your CRM and your inbox, so it can decide who to contact, write the message with real context, and send and log it on its own. Outsales, for example, scores intent on every contact and runs the follow-up end to end, which is the step ChatGPT leaves to you. The two are complementary: ChatGPT for the thinking and drafting a rep does, a connected tool for the execution. Our guides to AI sales tools and AI sales agents cover that second category.

The takeaway

ChatGPT earns its place in sales as a fast, cheap assistant for research, writing, objection prep, and summarizing. Give it real context and it saves a rep real time. Ask it for a cold email from nothing and it gives you the same template as everyone else.

Respect the limits. It invents facts, does not know your pipeline, and does not act, so it belongs next to a salesperson's judgment, not in place of it. Use it for the work around the conversation, verify what it tells you, and let a connected tool handle the part that has to actually happen.

Frequently asked questions

How do salespeople use ChatGPT?

Most use it for pre-call research, drafting and rewriting emails, practicing objection handling, and turning rough call notes into clean summaries and next steps. It is strongest as a drafting and prep tool when given real context, and weakest when asked to write outreach from nothing.

What is the best ChatGPT prompt for sales emails?

Give it a role, the prospect's context, and a specific ask, then request options. For example: act as an SDR, here is the prospect's background and recent company news, draft three short cold-email variants that reference the news and ask for a call. Context is what separates a useful draft from a generic one.

Can ChatGPT replace a sales rep or a CRM?

No. ChatGPT drafts and advises, but it cannot access your pipeline, decide who to contact, or send and log a message on its own, and it sometimes states facts that are not true. It complements a rep and a CRM rather than replacing either.

Is it safe to put customer data into ChatGPT?

Treat it with caution. Avoid pasting confidential or regulated customer data into a public AI tool, and follow your company's policy. Use anonymized or public information for research and drafting where you can.

What is the difference between ChatGPT and an AI sales tool?

ChatGPT is a general assistant you operate by hand for research and writing. A sales-specific AI tool or autonomous CRM is connected to your CRM and inbox, so it can act: choosing who to contact, writing with real data, and sending and logging the touch. They work best together.

Written by

Daniel Hayes

Revenue Operations

Daniel works at the intersection of sales and systems. He writes about CRMs, pipeline hygiene, and the workflows that keep deals from slipping through the cracks.

Related articles