ChatGPT for the sales department: 30 conversation scenarios

Why sales correspondence "drifts"

A familiar picture: in the morning a manager replies to a client in detail and warmly, by evening — dryly and in two words. One salesperson writes "Good afternoon!", another — "Hi". One handles the "too expensive" objection through value, another just gives a discount. Conversion fluctuates not because the managers are bad, but because there is no single standard for messages.

In a sales department, correspondence accounts for 70–80% of touches. A cold email, answering a question, handling "I'll think about it", sending a proposal, a reminder, an upsell. Each stage requires its own logic and tone. When a manager writes "whatever comes out", they reinvent the wheel every time — and spend hours doing it.

ChatGPT won't replace a salesperson. But it can become a "co-pilot" that drafts a message for a specific situation in 30 seconds. Below — how this works in practice and which sales prompts actually work.

What exactly can be automated

Let's break down the key steps of correspondence where AI delivers the maximum effect.

1. The cold first message. A manager spends 15–20 minutes researching the client and another 10 — on the text. With a prompt, they get 3 email variants for a segment in a minute. The main thing is to provide the inputs: who the client is, what hurts, what we offer.

2. Answering an incoming question. "How much does it cost?", "Do you have an integration with X?", "Send me the terms." ChatGPT formulates a reply in the right tone and with the right emphasis — not dry, but without fluff either.

3. Handling objections. "Too expensive", "We already work with others", "I'll think about it", "Not now". This is the most painful stage. Objection-handling prompts let you prepare 5–7 response scenarios in advance and adapt them to the client.

4. The commercial proposal. Not a 10-page template, but a short proposal tailored to the client's task: pain → solution → benefit → next step. ChatGPT assembles the structure and text, the manager corrects the numbers.

5. Reminders and follow-ups. "Soft" follow-ups that don't annoy but bring the conversation back. AI is especially useful here — it doesn't get tired and doesn't slide into "so, what have you decided?".

6. Upsell and cross-sell. After closing a deal — an offer of an adjacent product. The prompt helps formulate the value without being pushy.

7. Internal communications. A call summary for the CRM, a note for the manager, a brief for handing the client over to support.

Each of these steps is a separate correspondence scenario. If you assemble them into a system, the department gets a reproducible standard, not a set of improvisations.

The solution step by step — how it works

Technically, everything is simple. There's no need to program anything.

Step 1. Build a prompt base. Each scenario is a text template with variables. For example: "You are a sales manager for [product]. The client is [description]. They wrote: [message]. Reply in a [tone] tone, take into account the objection [objection], suggest the next step." Such a template can be reused dozens of times.

Step 2. Set up the context. The more inputs the manager provides, the more accurate the response. A good prompt always contains: the AI's role, the task, inputs about the client, constraints (length, tone, what must not be said) and the output format.

Step 3. Embed it into the daily routine. Opened ChatGPT → selected the right prompt → inserted the client's data → got a draft → refined 2–3 phrases → sent. The whole cycle — 2–3 minutes instead of 15.

Step 4. Make the templates shared. Prompts are stored in a single document or team note. A new manager doesn't spend months learning — they take a ready-made AI correspondence script and immediately work at the level of experienced colleagues.

Step 5. Update based on feedback. Once a week the team notes which scenarios worked and which didn't. The prompts get revised — the system becomes more accurate.

If you don't want to assemble 30 scenarios by hand, there's a ready-made set: ChatGPT for the sales department: 30 ready-made scenarios for correspondence and objection handling. There, the prompts cover the entire cycle — from the first touch to the upsell, with examples and explanations. It's faster than writing from scratch.

What the business gets

The main result is stability. Correspondence stops depending on the mood and experience of a particular manager. The client gets an equally high-quality response at any stage.

Next — time savings. In most cases, a manager spends 5–15 minutes on a single message. With prompts — 2–3. Across 20 messages a day, that's 1–2 hours of freed-up time. Over a week — almost a full working day.

The third effect — conversion growth through the quality of handling. Objections aren't ignored, follow-ups aren't forgotten, the proposal goes out the same day. Usually it's precisely at these stages that the most deals are lost.

The fourth — fast onboarding. A new manager gets not "read the scripts", but a working tool that suggests phrasings in real time.

The fifth — a unified tone of voice. All messages sound like one brand, not like five different people.

If you need not only sales but also marketing, content and routine — take a look at 50 ChatGPT prompts that save 10+ hours a week. It's a set of proven templates for various tasks, each with an example and instructions.

And if you need the widest possible base — Ultimate AI Prompt Library contains 2114 curated prompts and 12 workflows. This is no longer a list, but a ready-made system.

Where to start

Don't try to implement everything at once. Start with three steps.

1. Choose the 5 most frequent situations. Usually these are: a cold email, a reply to "too expensive", a reply to "I'll think about it", sending a proposal, a follow-up. This is where AI delivers a quick effect.

2. Write one prompt for each. Use the structure: role → task → inputs → constraints → format. Test it on real conversations.

3. Make the templates available to the team. One document, a link in the pinned message, a short call to walk through it. Within a week, managers will start suggesting new scenarios themselves.

Then expand the list: add stages, client segments, types of objections. Gradually the department will build a full-fledged library of sales prompts — and correspondence will stop being a lottery.

Conclusion

ChatGPT doesn't sell instead of the manager. It removes the routine of phrasing and gives the team a single communication standard. 30 correspondence scenarios are not about "AI magic", but about discipline: each stage of the sale gets its own high-quality text, regardless of who is on shift today.

If you want to not assemble prompts one by one, but immediately get a ready-made set for the sales department — start with the collection of solutions for Business. There you'll find everything you need to implement the system in a day, not a month.