The best way to use artificial intelligence to improve cold email conversion rates is to build better lead lists.
Most of the AI-enhanced emails that I receive these days have clearly used artificial intelligence to do research on my business, but they don’t consider whether or not my business is a good fit for them.
I get messages all the time that say “I see that Hacked BD does cold email, we could help you find more clients that want to send cold email”. We obviously don’t need outsourced business development considering that’s the service that we provide. Anyone that sends this message to me is wasting an opportunity that could have been used on a real prospect.
This article will go into more detail on the process that Hacked BD uses to personalize cold email campaigns for our clients with artificial intelligence. This should be applicable regardless of the tools that you use.
If you are already doing cold email marketing you should consider how much improvement is possible by figuring out what percentage of your leads are a good fit for your solution. Most companies will build an email list using very simple filters like: software company, 10-50 people, CEO, USA. That will only get you so far.
Is every single software company CEO actually a good fit for what you do?
The best cold email agencies will make sure that the prospects in your lead list have a good chance of needing your solution. Removing people that aren’t a fit is going to be the easiest way to improve open rates, reply rates, and interested rates.
You need to use standard filters to build your initial lead list, but you should use artificial intelligence to refine it.
Let’s use the example from above - you know that software companies in the US are a good fit for you, you want them to have 10-50 employees, and you target CEO’s. Use those filters to build your starter list.
Then ask yourself what actually makes these companies a good fit for you. Look at your customer base and figure out what attributes they typically have.
They may all be well funded, be in growth mode, have a solution cost of $500+ per month, be in a highly regulated industry, etc. Even if these things are harder to justify like “they are open to new technology” you should add them to your ICP. Start out with a detailed description of your potential customer and see if you can find information that will allow your AI to filter by those criteria.
Almost every company has information in multiple places across the internet. This could be social profiles like LinkedIn, review websites likeG2, search engines like Google Maps, lead databases like Apollo, or a number of other sources.
It takes longer to aggregate information from multiple sources which why people skip this step, but it will make your cold email campaigns more effective.
The best enterprise sales reps conduct this kind of research before approaching an account. They likely have a list of 100-500 companies that they are responsible for and their quota needs to come from that list. They don’t have the luxury of shotgunning emails to tens of thousands of prospects to get a lead, so they make sure each prospect is personally targeted.
Find a process to aggregate the data sources that contain the information you need and you will be able to personalize your email campaigns the exact same way.
Once you have all of the information about your prospects in one place you can use artificial intelligence to improve your lead list quality. You need to write effective prompts that tell the artificial intelligence how to interpret the information.
I like to describe AI as smart dumb technology. It can execute specific tasks like a human, but it isn’t smart enough to reliably execute generic tasks.
Box the AI in as much as possible to make sure that you get the answers that you need. Use questions that can be answered yes/no wherever possible.
The team at Hacked BD will always review a list of AI responses once they’ve been generated. This is a good practice for you to follow as well.
9/10 times the AI’s response will be correct, but we don’t want to act on bad information or say something that is incorrect on behalf of one of our clients. It is much faster to generate hundreds (or thousands) of lines of research and have a human validate the information than to have a human manually connect all of the research.
You now have a list of prospects that are actually a good fit for your solution. You will already benefit from higher email conversion rates, but you can improve them further by using AI to write personalized emails.
The most important thing to keep in mind is to keep your email copy believable. If you wouldn’t say something to a person in conversation don’t write it in a cold email template.
We will use AI generated copy as a supplement for our templates. Don’t go overboard to the point where it’s obvious that you generated the email with AI and automated the sequencing.
The best way to do this is to come up with custom variables for the AI responses that are part of a broader template. This way the AI only has to come up with a short sentence or two, and you can make sure your cold email copy still flows well.
