Know What’s BETTER than ChatGPT?  ChatGPT plus Experience. 

A detailed example of using AI to help with cold-prospecting emails

If you haven’t heard of ChatGPT by now, then you’ve probably been living in a cave somewhere going through a social isolation experiment.  Or uber-binging on Netflix specials.

I’ve been using AI to assist in my client projects, for almost 10 months.  Notice the operative word, “assist”.

Some marketers, salespeople and consultants claim they have AI write everything.  They claim it’s reduced their costs, and made them more effective in their jobs.   

While AI tools can be incredibly helpful creative assistants, I believe it’s a mistake to entirely rely on the technology of an LLM (large language model) to create sales enablement tools. 

On the other hand, AI provides real benefits, if used wisely.

Let me explain.

How They Work

LLM’s like ChatGPT use a lot of published information (mostly from before Sept 2021) to “predict” an answer that likely answers your query.  I’ve found that:

  1. The answers aren’t always relevant to your current situation today, and
  2. Specific solutions are based on data that may or may not be proven to work

You need to exercise thoughtful judgement over what the technology creates.  Otherwise, you risk putting out material that simply misses the mark. 

What worked 5 years ago, frequently doesn’t work in today’s business climate.  And if ChatGPT scrapes “solutions” from sources that gave bad advice, then the AI solution will also be bad.

That’s why AI can be used for research (with fact checking), idea generation, brainstorming, and creating drafts.

But it’s critical to fact-check… look at the original sources to check the authenticity and authority/expertise.  As well, put AI-suggested copy through your experience filter, to make sure it’s relevant to TODAY’s problem/solution presentation techniques. 

Let’s get specific, and look at the application of AI to creating a cold prospecting email.

The Good News

In a word:  speed. 

With a well-crafted prompt, AI can create an e-mail outline and fill in the details, very fast. 

With some tweaking, you can have a cold prospecting email that reads quite well, sounds conversationally businesslike, and incorporates generally recognized components, like proof builders and a call to action.

The Bad News

In a phrase:  Questionable effectiveness. 

Unfortunately, you’re never sure if AI is pulling it’s “recommendations” from an authoritative source that has been proven to work, or a source that just claims to be effective. 

How many times in real life have you received recommendations from someone that sounds really convincing, but in fact is just plain wrong?  Same thing applies with AI. 

Layer on the fact that AI advice is almost 2 years old, and you could be getting some advice that just isn’t effective anymore.  If it ever was effective. 

Note that I’m saying “could”.  If you’re inexperienced in cold-emailing techniques, you just don’t know.  But if you have some experience, you can take what AI pumps out and modify it in a way you know would be more effective

You can have the best of both worlds… fast results that have a good chance of working.

Examples

Let me show you what I’m talking about with a real example. 

I’ve produced a series of How-To articles over the last year or so, on producing effective prospecting B2B emails. 

Now, let’s see what ChatGPT-4 comes up with.

Here’s the prompt I used: 

“Let’s assume the product is called a widgetX. The target would be an operations manager in a manufacturing plant. The widgetX is a mechanical gear box used on mixing vessels in chemical processing plants, and it should reduce operating costs by 30% over a 2 year time period, over the next best competing product. Those savings are mostly due to lower power consumption and lower maintenance costs, according to recent studies at a good university. However, it’s also 20% higher in price, and is relatively new on the market so not entirely trusted yet. It comes in blue, white and black. It weighs 203 lbs, vs 350 lbs for the next main competing product, and is made in Israel but sold around the world. We’re selling this to a target in the USA. Create an outline for the cold prospecting email, in bullet form, complete with suggested text/copy for each section. Then come up with your recommended subject line.”

And here was the response, or output:

**Subject Line:** “Unleash 30% Operating Cost Reduction with WidgetX: A Revolutionary Gearbox Solution”

Dear [Operations Manager’s Name],

I hope this email finds you well. I’m [Your Name], an Account Manager with [Your Company], a pioneer in advanced mechanical solutions.

We understand that managing operations in a manufacturing plant is a challenging task, particularly when it comes to the efficiency and reliability of mixing vessels in chemical processing. High energy consumption and frequent maintenance often skyrocket operational costs, creating an unnecessary burden on your resources.

Given these challenges, I wanted to introduce you to WidgetX, our latest innovation. This mechanical gearbox for mixing vessels is specifically designed to significantly reduce operational costs. WidgetX’s state-of-the-art design guarantees a 30% reduction in operating costs over a two-year period compared to current market alternatives. This is due to its superior energy efficiency and low maintenance requirements, as confirmed by recent studies conducted by [University’s name]. Moreover, despite being 20% higher in price initially, the cost savings over time provide a solid ROI, making WidgetX a sensible long-term investment.

Weighing only 203 lbs, WidgetX is almost 42% lighter than its nearest competitor, ensuring easier handling and installation. It’s also available in blue, white, and black to seamlessly integrate with your plant aesthetics.”

Although WidgetX is a newer product on the market, it is manufactured in Israel, a country known for its advanced technological developments. We’re confident in the product’s ability to deliver on its promises, as it has already received positive feedback from initial users.”

I would love to arrange a call or meeting at your earliest convenience to discuss how WidgetX could benefit your operations specifically. I believe our product can bring significant positive changes to your manufacturing process.”

Thank you for considering our solution. I’m looking forward to hearing from you.

    Best regards,

    [Your Name]

My Analysis

On first glance, it’s a decent email.  It’s “by the book”, as in textbook driven.  Problem solution style.  Cool, right?

But will it get a conversation going, start a relationship and get to the next step with the prospect?  It’s not BAD, and in my experience it could produce some results. 

But it misses some fundamental cold-prospecting qualities that reduces it’s effectiveness.  Here’s my break-down of the email:

  1. Tone:  almost Conversational… not bad, but sounds a bit stilted, or formal.  It needs to be more like you’re talking to a business associate at a conference, or perhaps over lunch.
  2. Length:  It’s way too long.  The prospect is likely to tune-out before getting halfway through.  150 words or less is the target.  This email is almost 300 words.
  3. Subject Line:  Not too shabby.  It’s starts with a major benefit. Puts the benefit first, that’s good.  However, it comes across as “marketing speak” with a bit too much “loud” salesy language.  Also, using the target’s name has been shown to improve open rates.
  4. Opener:  targets don’t really care about you, especially initially.  The email needs to dive right into something personalized, about the target.
  5. Pain point identification:  This is done quite well, although it’s too long.  Get to the point faster.
  6. Solution introduction:  Done well, but too long.  And initial emails shouldn’t go into the laundry list of all features and benefits.  Stick to ONE benefit.  Don’t try to overwhelm with all details in the initial email, and do NOT try to deal with objections like it isn’t built in North America (wait until some future meeting).
  7. Proof:  stating the university name is good credibility, but there needs to be something stronger.
  8. There’s no interest-building question.  Questions are so powerful in relationship building.
  9. The CTA (a meeting) is both weak, and likely too big a step at this stage.  How about a more detailed fact sheet?  That one is easy to say “yes” to.  Although, it does a good job of selling the benefit of the suggested meeting.
  10. In general, cold emails need to be less like a marketing sell sheet, and more like a laser-focused “what’s the big thing that’s in it for me” message.  Remember, these people are BUSY! Once you get some traction in the relationship, then go into more detail. 
  11. The AI generated email actually reads at a grade 13 level.  That’s bad for comprehension of the average prospect.  Reading levels should be grade 7 or less. 

Modified Version

So here’s how I would change this AI output, taking my suggestions into account:

SL: [First Name], get 30% lower operating costs on your mixing vessels

Hi [First Name]

I noticed you’re in charge of operations at [company name].

If you’re like other operations managers we deal with, escalating operating costs are a real problem in your plant.

If that’s the case, you might be interested in our WidgetX mechanical gearbox.  Studies at [university name] show a 31% reduction in operating costs, and explosive slurry plants like yours in England, Israel and Spain have invested in our product to great success.

Would you be interested in finding out if you can get similar decreases in operating costs?

If so, how about we set up a short zoom call where I can show you those case studies, and you can see exactly how this would work in your plant.

I’m free next Tuesday and Wednesday, 10 am and 2 pm.  Get back to me with the time that works for you.

Sincerely,

[Your Name]

Do you see how much more focussed that is?

Notice how I inserted more “proof” or belief-building information?  You need to know when to do that.  Our AI friend didn’t think that was necessary. 

In Conclusion

In conclusion, AI can create a good draft that you can use to get to a final product.  It accelerates the process for you.  But you need to edit, restructure, and refine the output before using it, according to current best practises that have been shown to work. 

You might be wasting your time if you just entirely rely on AI.

In short: AI plus Experience = Results


For a deeper dive into the finer details of Sales Enablement tools and especially cold prospecting emails that work in today’s business environment, check out other articles on my website, and on my substack newsletter

If you’d like to receive future newsletters with even MORE practical tactical advice, consider subscribing to The Working Word, here

When you’re ready…if you’re interested in a private training session on how to get effective engagement with your prospects and clients, drop me an email at [email protected].