Generative AI Prompts for Sales

If you’re not using generative AI to maximize your selling impact, you’re behind the curve. In a recent report on AI trends, more than 80% of salespeople who use generative AI are creating basic content, like sales pitches and emails, and more than 70% said they’re using it to analyze market data.

But there’s a catch. Tools like ChatGPT for sales can deliver big productivity wins for reps, but it isn’t always easy to get exactly what you need. Knowing what to ask in a query — known as a prompt — is an art form in itself. You need to know your target audience, consider the context of the communication, put in effort to personalize and polish your messages — and always, always filter out confidential information.

Keep in mind — many customer relationship management (CRM) tools are adding built-in generative AI capabilities, so you may not even need prompts in the future. For now, I’ll share some of my favorite prompts that are ready to use, and offer guidelines for safe and ethical prompt use.

Attend the sessions, demos, and hands-on workshops that fit your unique role and business goals.

Generative AI creates content, like copy or images, based on simple, text-based prompts. There are many tools that allow you to do this, like ChatGPT, Bard, and Bing, but they all work in roughly the same way: A user inputs a prompt requesting specific content that doesn’t yet exist — like an email draft to a prospective customer. They set parameters, like how long they want the email to be and what they want it to include. Then, the generative AI tool finds elements of publicly available data that match the parameters of the prompt and pieces them together to create the desired content.

This is just a broad-strokes overview. You can read more about how this works in Salesforce’s article on generative AI for sales.

Generative AI has the potential to increase seller productivity, but it also introduces security concerns. In the same report cited earlier, about 73% of sales professionals said they’re worried about security risks with generative AI, and 49% don’t know how to use it safely at work. On top of this, generative AI often lacks enough context or background to create accurate responses, also known as “outputs.” So before we dig in, a few notes of caution on using this new tech:

Public generative AI tools are not secure. When you input data into tools like ChatGPT, you feed that information into the digital ether. It isn’t encrypted, leaving it vulnerable to unauthorized access and misuse. Exercise caution and refrain from sharing sensitive data such as passwords, financial information, or any personal information.

Outputs may contain inaccuracies. Public generative AI tools don’t have access to the most up-to-date information (ChatGPT’s latest dataset is from 2021). Also, these tools don’t vet information; they pull from a wide swath of online sources, some of which may contain incorrect information. Before sending any messages created using public AI tools, make it a habit to review the outputs for inaccuracies.

AI lacks human contextual knowledge. Generative AI, by its nature, is incapable of human emotion. It’s also not always able to connect the dots between different pieces of information, identify cause and effect, surface motivations, or use the most contextually appropriate language. (A recent sales conversation a Tesla AI salesbot had with a prospect shows this in spades.) Always review outputs to make sure the language is natural, includes sufficient context, and is tone-appropriate for the medium and audience.

Public generative AI tools lack customer context, too. While generative AI lacks human emotion and struggles to connect information dots, many CRMs now boast AI tools with built-in prompt functionality that pulls from customer information to provide more personalized and relevant outputs. They excel in recognizing cause and effect, understanding motivations, and can even mirror the voice of your customer based on previous communications stored in your CRM.

In my experience working with sales teams, these three applications of generative AI save the most time while increasing prospect and rep engagement: creating essential sales materials, like pitches and emails; drafting training content; and conducting basic research. The prompts below are the ones I see used consistently.

My recommendation: Have your senior reps experiment with these and fine-tune until they deliver impactful outputs easy to personalize. Once approved by managers, these prompts can be turned into templates for the entire sales team.

Example prompt: Create a presentation outline that includes {your company’s} mission statement, {your company’s} target customer, the benefits of {your product(s) or service(s)}.Prompt: Write a 1,000-word guide on {your product or service} including three statistics (with source) with 10 ways it/they will help {a specific role or buyer}.

Prompt: Describe the key factors that influence customer purchasing decisions for {product or service type} in {industry}.

Prompt: Summarize the top three critiques in customer reviews for {your product or service}.

Prompt: List the top three challenges for businesses in {target buyer industry}.

Prompt: Write a 50-word LinkedIn message explaining how {your product or service} can address the challenges of {a specific role or buyer} in {industry}.

Prompt: Write a friendly and engaging {social media platform} comment in response to a message that says {copy of recent social media post}, subtly introducing {your product or service} and offering a unique value proposition.

Prompt: Write a 30-second cold call sales script, highlighting three benefits of {your product or service} for {prospect/prospect’s company}.

Prompt: Write a three-minute video script for {title} at {prospect company} summarizing the benefits of {your product or service}. End the video script by reminding them of an upcoming meeting on {date and time}.

Prompt: Act like a {title} in {industry} who is hesitant about the cost of {your product or service}. Roleplay a conversation discussing the value and benefits the product/service offers.

Prompt: Act like a {title} in {industry} who is happy with {competitor solution}. Respond to a sales pitch that summarizes the value of {your product or service} and offers discounts for new buyers.

Or skip the prompts altogether and consider opting for a CRM with generative AI capabilities built in. That way you can generate responses not just from public databases, but also from your trusted customer data. The result? More personalized and relevant outputs without any prompt engineering. Here’s a demo of how Salesforce does this.

AI can log your activities, target the right leads, and boost productivity. Discover how on Trailhead, the free online learning platform from Salesforce.

Generative AI’s “open field” format has its pros and cons — you can create prompts for just about anything, but that doesn’t mean you’ll get something useful back. The quality of the response heavily depends on the quality of the prompt. By using a built-in AI tool in your CRM like Sales GPT, you can skip the guesswork and avoid tool-toggling by using out-of-the-box prompts tailored to top sales process use cases. But if you’re using a public generative AI tool like ChatGPT for sales, I recommend being direct and linear with your prompts. Here are some basic guidelines:

Use clear, specific language. Add numbers and dates, concrete adjectives, and any measurable parameters to ensure your prompts generate useful copy. For example, instead of prompting ChatGPT with “Tell me everything I should know about cars,” push for specifics: “Write three short sentences summarizing the auto industry’s biggest financial problems over the last 10 years.”

Keep it linear. Generative AI tools like ChatGPT like logic and flow. When crafting prompts, frame your asks in a linear format: “do this, then that, then that.” You can even use bullet points to frame the ask and make your request clearer.

Add context and background for more relevant outputs. If you want outputs to address specific information, note that information first. Add in email or social copy ChatGPT can reference, or, if information is publicly available, note that in your prompt. For example, instead of writing, “Draft a cold email to buyers in the automotive industry,” write “Draft an introductory email to the CEO of Tesla that shows how (your product or service) addresses the top three challenges of businesses in the automotive industry.” This happens automatically in Sales GPT, as every email generated is based on contextual CRM data.

Never forget personalization. ChatGPT doesn’t have access to personal information — and personalization is how you land prospects, close sales, and generate loyalty. To ensure you get that personal, hyper-relevant hook, make sure you add in the right names, titles, and account-specific details to any generative AI outputs. Pro tip: Sales GPT can pull straight from your CRM data to personalize communications with a click of a button.

Disclose the use of AI, when necessary. In situations where prospects would reasonably assume a human created the content, be sure to note generative AI tools were used. I like to include these as footnotes on presentations or in small print on email signatures whenever AI copy is added. Doing so will help protect your reputation and build trust.

Once you’ve picked the prompts that work well for you, test and refine them to improve the outputs — then personalize for that human touch. Also, remember: ChatGPT for sales is not a replacement for relationship-building. It’s a time-saving tool that gets you to meaningful conversations faster. Sales will always be about human relationships — and even the most advanced GPT prompts are unlikely to change that.