Prompting & cold emailing with AI

Table of contents

1. How does AI actually work?
2. What are Prompts?
3. Use Cases for AI in Sales
4. Prompt Engineering: How you can improve prompts
5. Bonus: Scale ChatGPT Tasks with Cargo AI and Google Sheets
6. My Experience and Conclusion

Are you looking for new ways to take your sales game to a new level? Then you’ve come to the right place!

In this article by our guest author Philipp Käming, it’s all about cold calling with AI and how you can smartly use artificial intelligence in sales. Whether you’re an experienced sales professional or just starting out, there are always ways to improve your sales strategies.

In this article you will learn what AI actually is and how to use it in sales.

To this end, I will show you various use cases so that you get a better feel for how best to use AI in cold calling. Additionally, we’ll take a look at how you can get AI text generator tools like to write strong cold mails for you … and how you can execute these requests 100-fold in one go to save even more time.

But before we get into the details, let’s first clarify what AI actually is.

How does AI actually work?

AI stands for artificial intelligence (English AI, Artificial Intelligence) and describes a technology that enables machines to develop human-like intelligence and perform tasks independently.

There are different types of AI. This includes machine learning, deep learning, and natural language processing.

Machine learning is a technique that allows computers to learn from data. Deep learning goes a step further and uses neural networks to recognize complex patterns. NLP enables systems to understand and process human language. Thanks to the training, the AI learns how to perform certain tasks and eventually makes independent decisions. Chatbots, like ChatGPT, can recognize patterns and correlations in data and generate statements based on them.

One of the important aspects in the use of tools like ChatGPT is the prompts.

What are Prompts?

Prompts are short sentences or phrases that serve as input for AI systems. They are essentially the task you give the AI. You can formulate them as instructions or questions and the AI will do the job. Prompts can also be loaded with a variety of sources, including examples, websites, or LinkedIn social media profiles. It’s important that you tell the AI exactly how it is to behave and what it is to do.

An example of a simple prompt in sales could be:

The result, which ChatGPT generates, looks like this:

Let’s stay in the same chat and go one step further. I’ll tell ChatGPT a little more about myself and ask for ways I could collaborate with OMR.

Let’s see what ChatGPT suggests to me:

Ok. Strong. It ended up being number three.

Now that you have an idea of how AI works and what prompts are, let’s dive into various use cases for AI in sales. Tools like ChatGPT can do a lot more than generate suggestions for cold mails.

For this, I’ll open a new chat with ChatGPT.

Use Cases for AI in Sales

Identify Persona with AI

One of the strengths of AI in sales is that it has learnt from a vast amount of data. This also means that it can interpret different user groups and their challenges. Based on your offer too.

Even if you already know your persona, ChatGPT can give you very exciting input, for example for opening up new target groups that you hadn’t thought of before.

Let’s ask ChatGPT:

And ChatGPT … what do you say?

Result ChatGPT Persona 1

Result ChatGPT Persona 2

Result ChatGPT Persona 3

This information can be used to create targeted campaigns and acquisition adapted to the needs of the leads.

The marketing manager doesn’t really fit for me. But Sales Manager and CEO are well defined. I decide for the CEO.

However, I need more information about the suitable companies so that I can also filter them out well and don’t send anything generic to “B2B companies of various industries”. After all, a high degree of personalization is crucial in acquisition and this is precisely where AI can come into play.

Let’s give ChatGPT the next task:

The result:

Top! I can work with that. My target group will be:

CEOs of B2B companies from the technology sector with 50 employees and €5 million in turnover. Their goal is to increase their revenue through new customer acquisition. The companies/CEOs are active on LinkedIn.

Ps. Of course, anyone who wants to level up their sales game can participate in my workshops 😉

So let’s stay in the same chat and tease out a few pain points for this persona. Of course, we stay in the same chat again, because ChatGPT also takes into account the previous history.

Identifying Pain Points with AI

For successful acquisition, we need to put ourselves in the recipient’s shoes and understand their needs and pain points. A pain point is a problem or challenge that the customer wants to solve.

It can be difficult to identify these pain points, especially if you know little about the customers. But AI can help you to collect this information.

Let’s ask ChatGPT about Pain Points:

The result:

ChatGPT delivers very good pain points here, which just about every CEO faces on a daily basis.

So we have a good basis for now. We know the persona with their goals and challenges, key figures of suitable companies and the pain points.

For maximum personalization, we still need the interests. For this, I have a LinkedIn profile of a suitable lead and … ask ChatGPT.

Finding Interests with AI

But why actually? If you can understand the customers’ interests, optimal personalization can take place. Personal interests without a business context are best suited for this.

But that also involves a lot of work. You have to scan websites and social media profiles to find the crucial information. But you can also feed ChatGPT, for example with the LinkedIn profile, more precisely with the “about me” area.

Let’s take a look at how exactly this looks like:

The result:

Ok. Example 1 is total demolition. Number 2 and 3 don’t quite fit, but one is enough.

Let’s build a LinkedIn invite from this… let.

Writing LinkedIn Invites with AI

LinkedIn is an important channel in B2B sales for making new contacts and networking. But LinkedIn is also ideal for cold calling. Especially in combination with strong cold mails, an unbeatable combo.

Also here applies: The better the prompts and the training material, the better the results. But since we now have quite a bit together, it should be no problem.

Let’s see what ChatGPT says:

The result:  

That I offer AI tools is not quite correct. But I offer the suitable strategy for it. Change one, two words and we have made a really good LinkedIn invite.

Writing Cold Emails with AI

Cold emails are an important part of B2B sales. However, they are very time-consuming and require a lot of effort and skill in order to achieve a high success rate.

Since we want to save time, we let the artificial intelligence write us a suitable cold mail directly.

The result:

Very strong. Here too, a few minor adjustments have to be made and then we can work with it.

In general, the better the training and prompts, the better the result. You should not rely on a single result here, but try out and test several variants in order to achieve the best performance.

It’s also important to mention here that you should never blindly send out the AI-generated results. Sometimes they are simply wrong or not suitable for what you need.

Prompt Engineering: How you can improve prompts

The quality of your prompts and the delivered input are, as I said, key to good results.

To ensure that your prompt works optimally, you should test with various inputs and optimize until you get your desired result.

With ChatGPT, you can also adjust your prompt via the symbol at the top right, ChatGPT also saves your previous prompts. This way you can compare and optimize well.

I have also collected a few more tips for you on how to continually improve your prompts.

Use Specific Information

Compile your prompts as detailed as possible. This increases the likelihood of generating results that match your ideas. For example, if you’re writing an AI email to potential customers, you should include information like the company name and the industry in which it operates. You can also use URLs from websites or LinkedIn profiles.

Experiment with Different Formats

There are many different types of prompts. For example, you can ask a question, make a statement, or even enter a keyword and your AI should generate a complete answer from it. A special trick here is to tell the AI how to behave.

For example like this:

• Behave like an email copywriter
• Behave like a content creator on LinkedIn
• Behave like a sales expert

Use Alternative Prompts

If you don’t get the results you want, test alternative prompts and check the results. Sometimes a small change can strongly influence the result. For example, you can try a different keyword or a different question.

Use Feedback to Improve

You can also give feedback to ChatGPT. Let’s take the cold mail to John as an example to illustrate this.

The result:

Much better. Especially when you consider that 20 minutes ago we didn’t even have a target group.

Bonus: Scale ChatGPT Tasks with Cargo AI and Google Sheets

We’ve been busy and have built a lead list. For example, there are 6 example leads in here. All perfectly match the personas and companies we want to address.

Ps. The step of creating the list can also be “significantly accelerated“.

So we have names, position, company name & website and LinkedIn URL plus “About me” description of the leads and the industry. In order to execute multiple requests in one go, we have to link ChatGPT with Google Sheets via CargoAI.

Link Google Sheets with ChatGPT via CargoAI

First we install the extension CargoAI and grab the ChatGPT API Key.

OMR GIF: Connection

And we are good to go.

Let The Interests Be Created

The Prompt:

Behave like a sales expert. Analyze the following “About me description” and find an interest that implies the same values as mine. In one sentence, describe why this is so.

My values:

• Creativity, precision, efficiency, and speed

The about me description:

GIF: Let Interests Be Created

Script Cold Mails

Behave like a sales expert. Use the following information to pre-write me a cold email that I can send to the lead.

Start with the shared values you found on LinkedIn. Then make a transition to my solution. The goal is to have a conversation about my further ideas. Write a maximum of 800 characters, be sympathetic and use informality.

About me:

My name is Philipp and I help companies with automated acquisition via LinkedIn and cold mails, I also know a lot about AI. With this combination, B2B companies generate high-quality leads and increase revenue without marketing costs.

About the lead:

• Name: {{Placeholder}}
• Company name: {{Placeholder}}
• Interest: {{Placeholder}}

GIF: Script Personalized Cold Mails

My Experience and Conclusion

With AI tools like ChatGPT, you can not only find inspirations for anything, but also save a lot of time. AI tools really have the potential to take your sales to the next level.

I myself frequently use AI tools in the area of acquisition and copywriting, often ChatGPT spits out results that I would not have come up with myself. Since I send out mails and LinkedIn invites regularly in my everyday life, I could test how much time you save and how performance increases at the same time. For example, my mails which are adorned with pain points, challenges, or also interests from AI research have a reply rate of up to 5 %, as generic templates.

With LinkedIn invites, I get, if there is something usable in the “about me” description, on sometimes over 30 % acceptance rate, which is very strong. If I did this completely manually, I would need 5–10 minutes per lead. If I do it with ChatGPT and CargoAI, it’s only 2–3 minutes because the messages need to be adjusted at the end of course.

In summary, I can only recommend that you try it out for yourself and let yourself be inspired. But here again the hint: Sometimes the results are simply plain wrong or completely inappropriate. Therefore, you should never use unverified AI content.

Recommend AI-Text-Generators

On our comparison platform OMR Reviews you can find more recommended KI text generators.
There are over 60 different systems to choose from, tailored to the specific needs of small and medium-sized companies, start-ups and large corporations. Our platform offers comprehensive support in all areas of text creation and optimization. Take the chance to compare different AI tools and consult real user reviews to find the perfect tool for your specific requirements:

Originally Appeared Here