We all know that artificial intelligence tools can be used for email campaigns. This technology is attracting the attention of many companies because it promises to increase efficiency in various marketing tasks.
But are expectations borne out in practice? How does the introduction of AI into the design of a marketing campaign impact performance? What is the best way to use algorithms in these processes?
I’ve wanted to investigate this subject in detail for some time, so I decided to dive into GetResponse’s database and see how our customers have been doing with AI lately.
As email marketing is the heart of our platform, I chose this strategy as a reference for my research.
From now on, you’ll see statistics on the use of AI in more than 16,000 emails sent within GetResponse by companies of different industries, sizes, and budgets.
Let’s compare the performance of messages created within our AI generator and those that users have created from scratch in the classic template editor.
How we selected the data
Before we get to the numbers, let me briefly explain the process we followed for this survey.
GetResponse is a global email marketing platform with clients in various market segments. At the beginning of last year, we incorporated an AI feature for writing and designing emails equipped with technology from OpenAI (the creator of ChatGPT).
This tool has not replaced our email and newsletter editor, where users can choose a template and write the message from scratch. Therefore, there are two methods available within GetResponse.
The AI email generator receives information about the user:
- Type of message (welcome, newsletter, etc.);
- The subject of the email;
- Type of business (gym, clothing store, restaurant, etc.);
- The email’s target audience (busy university students, recently graduated engineers, etc.).
Then, the client can choose between different tones of wording for the message, such as friendly, formal, convincing, etc. The next step is to define the layout and color scheme of the email.
With the draft done, the user can also rely on AI to create the message’s subject.
The method is based on the premise that AI can use all the user’s choices to “program” the ideal email for each campaign, considering that the client knows a lot about their offer and their audience’s behavior.
Our analysis aimed to determine how this concept is reflected in the results. We selected all the emails with some components created with AI (subject or text and layout) and sent in 2024.
We compared these messages with those produced using the classic method and analyzed the performance of the two groups on the following criteria:
- Open rate;
- Click-through rate (CTR);
- Click-to-open rate (CTOR);
- Unsubscriptions;
- Spam complaints.
In addition to the general statistics for these email marketing metrics, we have broken down the data by the market segments in which the clients operate. The idea is to find out which sectors have used AI the most and see if this technology has worked better for some than others.
Finally, we consider the data from GetResponse’s latest Email Marketing Benchmark, in which we evaluate our clients’ performance using different metrics.
The survey gives an overview of the application of generative AI in email campaigns. So, I recommend using this data only as a sample and not as a direct guide to making decisions in your strategy.
That said, let’s get to the statistics!
Which is better: Emails generated by AI or those created manually?
Let’s start with an overview of the average performance of each type of email using some important metrics.
Here is a quick explanation of what each one indicates.
The first three metrics in the table are directly related to user engagement because they indicate whether the user was interested in opening and interacting with the content and links presented in the message.
The last two metrics include aspects beyond the content’s quality. A high unsubscribe rate, for example, may indicate that the sender has collected a lot of email addresses without people’s permission, and they leave the list when they realize this violation.
| Metric | Generated with AI | Created from Scratch |
|---|---|---|
| Open rate | 37.37% | 41.05% |
| Click-through rate (CTR) | 9.44% | 8.46% |
| Click-to-open rate (CTOR) | 25.25% | 20.62% |
| Unsubscribe | 0.16% | 0.14% |
| Spam complaints | 0.01% | 0.01% |
This general comparison shows a balanced scenario in the results of the two types of emails with fluctuations in some criteria.
Open rate: Emails Created from Scratch performed better
Emails created without AI had a slight advantage in the open rate, and it’s difficult to attribute the difference to the absence of AI in the process.
We could even assume that the subjects of these emails are more attractive to users, but I’ll discuss that in more detail in a specific table.
The monitoring of open rates is influenced by many factors, such as:
- Segmentation of email lists: if we send offers to users who are not interested in them, they tend to ignore the message
- Privacy rules on some devices: Apple, for example, allows users to view email without opening it, which directly interferes with the analysis
- Sender’s Reputation: depending on the name of the domain sending the message and the way it appears in the inbox, the user may click to open the email or ignore it
- Time of sending: some specific times of the day may be more or less favorable for recipients to consume the content of an email
- And several others, such as the demographic profile of the audience, the type of email sent, the reader’s preferences, etc.
I wanted to highlight these points to show that your open rate won’t magically rise or fall if you use AI to craft your emails. But it’s an interesting experiment because the algorithm can help you define the message’s ideal audience, structure, and headline.
Then, refine the content and design to ensure consistency with your brand, shoot, and track openings. You can never test too much!
Click-through rate (CTR): AI-generated emails win
CTR is perhaps the most important engagement metric in email marketing. It shows the level of interest people have in the content presented to the point where they click on a link within the message.
In this dispute, the emails created with the support of AI won by a narrow margin. We can’t say it was the main reason for the result, but considering the CTR value, it’s a more significant difference than the open rate.
The number of clicks on an email can vary according to:
- The relevance of the content to the user: imagine that you lured the person with a subject that promised X but delivered an offer Y in the body of the email. In this case, it might be difficult to get that click, but it might happen if it meets the person’s expectations and captures their attention.
- The type of email sent: transactional messages such as “your order has arrived” tend to have much lower click-through rates than informative newsletters, for example.
- User experience: If the email doesn’t work well on the mobile screen or the link button has an error, the click-through rate can also be negatively impacted.
But how can AI help you get more clicks on your campaigns?
You tell the tool the subject of the email, the approach, the tone of voice, the company profile, and the audience. It uses this information to create an email proposal that meets the specified objectives.
Then, your job is to edit what the AI has delivered and include links strategically to encourage the click and take the user where they want to go.
It’s a process in which your participation is fundamental because the output quality depends on the accuracy of the instructions you give to the AI.
But it’s also worth testing the classic method without AI to compare different outputs and performance. Some campaigns are special and require more direct intervention from the brand’s manager to get the clicks that will lead to conversions in the future.
Click-to-open rate (CTOR): AI-generated emails win
The analysis of this metric depends on the previous two (Open rates and CTR) to make sense.
The general rule is that the higher the CTOR, the better because it considers the number of people who opened the email and those who clicked on a link within it.
CTOR was higher in the emails created with AI. Our guess is that the algorithm’s targeting of the audience and tone of voice may have helped generate more clicks even with fewer opens, making it more effective.
Emails created from scratch show lower unsubscribe rates
Retaining users on your email list is fundamental to building a solid community for your brand. And if you don’t prepare your messages, you could alienate your recipients.
In our survey, emails created without AI won by a small margin. In theory, this could mean that the messages made from scratch had a communication style that was a little more consistent with what the brand usually applies and didn’t negatively affect the user.
But just like the other metrics, the unsubscribe rate varies for different reasons beyond AI’s use in message construction. Some of these are:
- User consent: those who buy generic email lists, for example, run the risk of having very high unsubscribe rates;
- Frequency of sending: if the person feels that you are communicating too much, they can eventually press the unsubscribe button;
- Forced or misleading content: a message with a very aggressive commercial tone trying to push a product or misleads the user with clickbait generates many unsubscriptions.
If you want to use AI in your creative process, pay attention to the tone of the message to not scare the user with a confusing approach or one incompatible with your style.
Another essential action is to connect the email’s audience with the list of recipients who will receive it.
Imagine that you have a recipe newsletter for different consumer profiles, and you decide to use AI to create an email about the best meats for the frying pan, but you accidentally send it to the list of vegan users. Unsubscribing can taste even more bitter in this case.
Spam complaints: a general draw
One of the most common assumptions about generative AI is that the content it creates is inaccurate and unreliable. This claim is true in some cases, such as the screenshots of bizarre search results made with Google’s AI feature this year.
In email marketing, sending such messages can be fatal to the sender’s credibility, with the risk of receiving the dreaded spam reports.
They happen when the user clicks the “mark as spam” option in the inbox, which leads email providers to weaken the domain’s reputation.
Our survey indicates that emails created with and without AI had the same rate of spam reports (very low in fact thank goodness!) so we don’t have any indisputable conclusions here.
Here’s a reminder of the good practices I mentioned above. Segment your list well, carefully review the content of the email before sending it, share really relevant offers with your audience, and don’t force communication. That way, no one will have any reason to report your emails, whether they’re made with AI or not.
AI-generated email statistics by market segment
Has AI generated more encouraging results for email campaigns in specific areas? Does this technology work in the same way for all types of companies?
Now, let’s see what GetResponse’s customer data reveals.
Open rate
The average open rate of our Benchmark last year was 26.8%.
| Sector | AI-generated | Created from scratch |
|---|---|---|
| Agencies | 57.2% | 43.79% |
| Arts and Entertainment | 44.88% | 52.34% |
| Automotive | 38.98% | 36.65% |
| Communication | 49.60% | 62.92% |
| Education | 31.81% | 44.61% |
| Finance | 40.19% | 40.84% |
| Beauty and well-being | 46.11% | 41.44% |
| Health | 61.23% | 47.65% |
| Digital Marketing | 52.10% | 33.77% |
| Legal | 41.94% | 57.13% |
| NGOs and non-profit organizations | 57.94% | 59.88% |
| Editorial | 24.26% | 46.24% |
| Real Estate | 30.79% | 38.28% |
| Restaurants | 38.05% | 47.99% |
| Retail | 35.07% | 38.84% |
| Sports | 27.10% | 42.08% |
| Technology and IT | 48.08% | 42.86% |
| Tourism | 28.79% | 32.85% |
| Other (not informed) | 36.50% | 40.56% |
In general, emails created from scratch had better open rates in almost all sectors, especially Sports, Editorial, and Legal.
On the other hand, Healthcare, Digital Marketing, and Agency companies achieved excellent figures for emails generated with AI. In other areas, the results were more balanced, but in the end, the victory went to messages prepared without the use of generative algorithms.
Click-through rate (CTR)
In the latest GetResponse Benchmark, the overall average CTR was 1.89%.
| Sector | AI-generated | Created from scratch |
|---|---|---|
| Agencies | 29.05% | 12.98% |
| Arts and Entertainment | 29.00% | 13.64% |
| Automotive | 3.19% | 9.85% |
| Communication | 10.90% | 34.67% |
| Education | 9.95% | 8.57% |
| Finance | 1.23% | 10.80% |
| Beauty and well-being | 4.77% | 4.50% |
| Health | 14.23% | 6.45% |
| Digital Marketing | 18.65% | 3.97% |
| Legal | 22.87% | 30.47% |
| NGOs and non-profit organizations | 21.65% | 21.14% |
| Editorial | 1.13% | 12.70% |
| Real Estate | 9.42% | 8.97% |
| Restaurants | 2.14% | 7.80% |
| Retail | 9.47% | 15.71% |
| Sports | 79.36% | 22.59% |
| Technology and IT | 19.67% | 19.97% |
| Tourism | 20.04% | 6.07% |
| Other (not informed) | 7.53% | 7.73% |
It’s tough competition here—AI-generated emails were superior in 10 industries, and non-AI emails were better in 9. This is a reflection of the overall averages we saw above. Another result that had a big impact on the final tally was that of the Sports sector, which had an extraordinary CTR.
Some segments achieved much higher rates in emails created from scratch, such as Publishing (11x!), Finance (almost 9x), Automotive, Restaurants, and Communication (approximately triple). 4 of these 5 also had worse results with AI emails regarding open rates.
Other areas saw much greater engagement with the messages created with the support of our AI email generator. In addition to Sports, we highlight Tourism (triple), Digital Marketing (almost 5x), Health, Arts and Entertainment, and Agencies (just over double).
Click-to-open rate (CTOR)
Our clients’ average result in this metric last year was 7%.
| Sector | AI-generated | Created from scratch |
|---|---|---|
| Agencies | 50.73% | 29.64% |
| Arts and Entertainment | 64.63% | 26.07% |
| Automotive | 8.18% | 26.88% |
| Communication | 21.97% | 55.10% |
| Education | 31.27% | 19.21% |
| Finance | 3.06% | 26.46% |
| Beauty and well-being | 10.36% | 10.87% |
| Health | 23.24% | 13.53% |
| Digital Marketing | 35.79% | 11.76% |
| Legal | 54.52% | 53.34% |
| NGOs and non-profit organizations | 37.37% | 35.30% |
| Editorial | 4.67% | 27.46% |
| Real Estate | 30.61% | 23.44% |
| Restaurants | 5.63% | 16.26% |
| Retail | 27.00% | 40.45% |
| Sports | 292.84% | 53.67% |
| Technology and IT | 40.92% | 46.60% |
| Tourism | 69.61% | 18.48% |
| Other (not informed) | 20.63% | 19.06% |
Here, emails created with AI were better in 11 industries. The results are similar to the CTR table because of the direct relationship between the metrics.
We only noticed differences in the Legal, Beauty, & Wellness sectors. In the former, the AI-generated emails had a better CTOR; in the latter, the winner was the method of drafting from scratch but with minimal margins.
The Sports sector’s CTOR is striking, completely off the charts! This is possible because this metric is calculated based on the total number of clicks and opens. Therefore, an email can be opened more than once with several clicks, perhaps a good sign of engagement.
Unsubscribe rate
Our clients’ average performance in this metric last year was 0.1%.
| Sector | AI-generated | Created from Scratch |
|---|---|---|
| Agencies | 0.41% | 0.19% |
| Arts and Entertainment | 0.15% | 0.19% |
| Automotive | 0.28% | 0.14% |
| Communication | 0.18% | 0.10% |
| Education | 0.17% | 0.18% |
| Finance | 0.08% | 0.09% |
| Beauty and well-being | 0.27% | 0.16% |
| Health | 0.40% | 0.24% |
| Digital Marketing | 0.15% | 0.15% |
| Legal | 0.55% | 0.19% |
| NGOs and non-profit organizations | 0.24% | 0.14% |
| Editorial | 0.06% | 0.13% |
| Real Estate | 0.17% | 0.13% |
| Restaurants | 0.23% | 0.24% |
| Retail | 0.14% | 0.14% |
| Sports | 0.25% | 0.14% |
| Technology and IT | 0.14% | 0.14% |
| Tourism | 0.14% | 0.10% |
| Other (not informed) | 0.17% | 0.13% |
Following the general averages, emails created from scratch had the upper hand when it came to unsubscriptions.
It’s a historically low rate in all business strategies and areas, so we don’t see major variations in this comparison.
The Health, Legal, and Agencies sectors had worrying results in AI-generated emails.
We could assume that the first two exhibited this behavior because they deal with more sensitive topics where the dissemination of unverified information has a very negative effect on the user. But this hypothesis would only be confirmed if we could analyze the details of each campaign sent.
Spam reporting rate
Our latest Benchmark revealed an overall average of less than 0.01% for customers in this metric.
| Sector | AI-generated | Created from scratch |
|---|---|---|
| Agencies | 0.03% | 0.01% |
| Arts and Entertainment | 0.01% | 0.01% |
| Automotive | 0.04% | 0.02% |
| Communication | 0% | 0.01% |
| Education | 0.01% | 0.01% |
| Finance | 0.01% | 0.01% |
| Beauty and well-being | 0.02% | 0.01% |
| Health | 0.02% | 0.01% |
| Digital Marketing | 0.01% | 0.01% |
| Legal | 0.03% | 0.01% |
| NGOs and non-profit organizations | 0.02% | 0.01% |
| Editorial | 0% | 0.01% |
| Real Estate | 0.03% | 0.01% |
| Restaurants | 0.02% | 0.01% |
| Retail | 0.01% | 0.01% |
| Sports | 0.02% | 0.01% |
| Technology and IT | 0.01% | 0.01% |
| Tourism | 0.01% | 0.01% |
| Other (not informed) | 0.01% | 0.01% |
Despite the very favorable score for emails created from scratch, the variations in the results between the two methods were very small.
This behavior is also because spam reports require stronger user action, as not many people take the time to mark the sender as spam directly.
The Legal, Real Estate, and Automotive sectors achieved the worst rates, but they are not alarming.
Email subject line generated by AI or manually: which is better?
I also thought it would be interesting to include this comparison in the study because subject lines can directly and indirectly influence user interactions with email.
The headline’s creativity, emojis or pre-headers, and other ingredients are decisive for the performance of subjects and emails as a whole. After all, if your headline fails to capture the recipient’s attention, all your rates could suffer.
GetResponse’s AI subject line generator uses very simple logic, similar to the email builder.
You enter keywords that indicate the central theme of the message and your company’s segment. You can also choose whether the subject will have emojis. The algorithm cross-references these instructions and delivers a list of options you can edit or not use.
Has this tool brought good results for our clients? Let’s find out.
Although the subject line directly affects the opening of the email, I decided to specifically analyze the CTR because it is a more accurate and valuable engagement metric.
After all, if the subject line is clickbait, for example, and the content isn’t relevant to the user, the open rate may be high, but it won’t do much good if a low CTR accompanies it. The message and subject line need to be coherent with no loose ends.
| Sector | CTR of AI-generated topics | CTR of subjects created from scratch |
|---|---|---|
| Agencies | 30.59% | 12.98% |
| Arts and Entertainment | 41.10% | 13.64% |
| Automotive | 6.21% | 9.85% |
| Communication | 11.20% | 34.67% |
| Education | 10.24% | 8.57% |
| Finance | 1.17% | 10.80% |
| Beauty and well-being | 4.79% | 4.50% |
| Health | 14.49% | 6.45% |
| Digital Marketing | 17.59% | 3.97% |
| Legal | 29.65% | 30.47% |
| NGOs and non-profit organizations | 21.47% | 21.14% |
| Editorial | 1.13% | 12.70% |
| Real Estate | 10.10% | 8.97% |
| Restaurants | 1.26% | 7.80% |
| Retail | 9.42% | 15.71% |
| Sports | 83.04% | 22.59% |
| Technology and IT | 21.84% | 19.97% |
| Tourism | 20.14% | 6.07% |
| Other (not informed) | 8.19% | 7.73% |
Total: 10.09% vs. 8.46%
AI-generated topics achieved a slightly better overall result.
The Sports, Agencies, Tourism, and Arts & Entertainment companies had great numbers in the topics created by AI.
On the other hand, titles created from scratch did much better in Restaurants, Publishing, Communication, and Finance. The scenario is similar to the previous metrics!
25 ways to build your contact list
We’ve compiled a list of 25 tried-and-tested tactics for the success of your future campaigns.
In conclusion
Email marketing is a strategy full of nuances and criteria that is needed to generate positive results in all the important metrics.
The presence or absence of AI in email design is not enough to determine the success of your campaigns. However, the data from our survey indicates that it can be a good tool to help your company become more effective in its communication.
In some sectors, the prospects for AI seem brighter than in others and vice versa. But that doesn’t mean that the average behavior of our customers will be exactly the same in your case.
At the end of the day, your participation in the process is fundamental for emails to perform well, with or without the help of AI.
After all, it works according to your instructions, and you always have the option of modifying the emails it delivers to make them your own.
Even if the tool makes mistakes and creates versions you don’t like, consider it more of a support and planning resource than a piece of software that will do all the work for you.
Test both methods, try giving the algorithm different types of information to train it in your brand’s style, and vary the approaches to find out which works best for your audience.
Want to see GetResponse’s email creation tools in practice? Create an account on our platform and test all the features for 30 days for free. You don’t need to put down your credit card to get started. Get to work!