Using AI for email marketing: Dos and don'ts

Email marketing: go with audience segments or go home. The more personalization you have per segment, the more successful your emails will be... however, the more personalization you have per email, the more work you have.

Enter AI to save the day... or just a bunch of time. However, there are some marketing slip-ups that can occur when AI isn't used right.

Here are the dos and don'ts of using AI for email marketing content—are you guilty of any of the Don'ts?

✅ Dos: How to use AI for email marketing content

1. Personalized subject lines at scale

   
   So, you want to send personalized subject lines at scale? That's great... how about making your life easier and getting AI to help? Input specific insights and data about each audience, industry, or region you want to target and customize your content for, and let AI fly with it.
   
   Example prompt: "Generate 10 personalized subject line variations for our summer sale email that would appeal to different customer segments based on their previous purchase history: frequent shoppers, occasional buyers, and first-time customers"
   
   

2. Optimized email send times

   
   When you send regular emails, as well as a lot of different types of emails (eg. newsletters, outreach, product updates... the list goes on), you get a good amount of useful data back. The only problem is, data can take a lot of time—and a headache—for you to go through.
   
   Use AI to help you spend less time on trawling through datasets (and spend less money on painkillers) and report back with super helpful information like the time of day that sees the highest open rate.
   
   Example prompt: "Analyze our email engagement data from the past 6 months and recommend optimal send times for each segment of our audience, considering time zones, industry, and past open patterns"
   
   

3. Email copy variations

   
   Don't you just love it when you finally finished a great sequence of emails? But then, out of nowhere, a request appears...
   
   Alexa, play the Jaws theme tune.
   
   "Can we have an A/B test for every subject line, and try out different industry-specific content versus more generic in every single one of the emails, please?". It's okay, marketers—don't cry, AI can help you create multiple versions of your email copy to help with your testing needs. Just make sure you're editing or watching over the content that is being sent out; you don't want your emails sounding like they were written by a robot.
   
   Example prompt: "Create three variations of this product announcement email: one emphasizing time savings, one focusing on cost benefits, and one highlighting innovative features. Each version should maintain our professional yet conversational brand voice"
   
   

4. CTAs (that actually convert)

   
   Deciding which of the gazillion different CTAs you could choose in your emails is a lengthy process in itself. Make it a little less painful by letting AI digest your already-written contnet and make suggestions on a CTA that would work.
   
   But don't get too carried away—the last thing you want is for your CTAs to sound desperate or pushy. Keep it as natural and flowy as possible, and you'll (hopefully) see those click-throughs and conversions soar.
   
   Example prompt: "Suggest 5 high-converting CTA options for our webinar invitation email that go beyond 'Register Now.' Include both the button text and a brief line of supporting copy that creates urgency without sounding pushy"
   
   

5. Direction for your next drip campaign

   
   Drip campaigns are a great way to nurture leads, but we can all agree: they take up quite a lot of your time. Most of the time, drip campaigns will have one aim, which means writing the same (or similar) thing in different ways, multiple times. Save your brain from imploding and get AI on the case.
   
   Give your AI tool all the information you have, including a brief, target audience insights, tone of voice prompting, and so on. Why? Because if you keep it generic, they'll be unsubscribing faster than you can say "AI".
   
   Example prompt: "Design a 5-email nurture sequence for new subscribers who downloaded our UX design guide. Include recommended timing between emails, subject line ideas, main talking points for each email, and logical next steps that lead toward a product demo request"
   
   

6. Email accessibility

   
   AI can also help you make your emails more accessible, which is super important. You want to make sure your emails can be consumed effectively by everyone, including those using assistive technologies.
   
   Example prompt: "Review this promotional email and suggest improvements for accessibility, including alternative text for images, better color contrast recommendations, and more screen reader-friendly formatting"
   
   

7. Cross-sell and upsell opps

   
   With the right AI tool, you can make sure your email content is making the most of any cross-sell or upsell opportunities, without needing to get sales or product involved. Genius. Feed your AI tool with customer data, as well as product details, data, and additional knowledge to help it spot appropriate moments for a sneaky product recommendation.
   
   Example prompt: "Based on purchase patterns in our customer database, create an email template that dynamically recommends complementary products to customers who purchased our home office furniture, with separate recommendation logics for different price point buyers"

❌ Don'ts: How to NOT use AI for email marketing content

1. Using templates with zero customization

   
   Risk: Yeah sure, you can use AI to create email templates that are just ready to go when you need them, but you risk generic and sometimes repetitive customer experiences. This can lead to your audience disengaging in your email content, lowering your click-through rate, open rate, and ultimately, making your unsubscribe numbers rocket. 📈
   

   Solution: Always provide AI with specific brand guidelines, examples of successful past emails, and clear objectives. Make sure you're reviewing and refining all AI-generated content to ensure it aligns with your unique voice and campaign goals.

2. Forgetting we all appreciate that human-to-human interaction most

   
   Risk: Because we live in a world of AI hype, filled with AI-generated content, we are starting to know how to detect AI content pretty easily. This has lead to us craving that clear human connection more than ever; without it, content (including your AI-generated emails) because off-putting and not-so-fun to read.
   

   Solution: Use AI to handle personalization at scale, but ensure your emails still feel human by keeping a human editor in the email workflow. Focus on story-telling, employee perspectives, and genuine brand personality that AI alone cannot replicate.

3. Going along with misleading or clickbait subject lines

   
   Risk: 47% of people will open an email based on the subject line alone. When you're dealing with an average AI tool, you're most likely dealing with something that just crawls the internet for something relevant to say back to your prompt, lacking a true understanding of your brand, best practices, and your target audience. Without the right information, you could end up with clickbait subject lines that make your customers click 'Delete' on your newest email instantly.
   

   Solution: Train your AI on examples of honest, value-driven subject lines that have performed well, or maybe even ones you've liked from your competitors or personal emails. Establish ethical guidelines for AI use that prioritize transparency and customer trust over short-term metrics.

4. Sending excessive emails because AI makes it easier to create the content

   
   Risk: We are the last ones to complain about the benefits of AI, namely how it helps marketers' productivity... HOWEVER, just because you've saved yourself a bunch of time on your email marketing content by introducing AI does not mean more emails. No customer wants to be spammed.
   

   Solution: Use AI to optimize your email cadence based on engagement data, not just to produce more content or create content faster. If you haven't already, you should have frequency caps and preference centers to respect subscriber boundaries. Long story short—don't overdo it.

5. Ignoring deliverability factors when optimizing for engagement

   
   Risk: There are a bunch of metrics that you can use to track the success of your team's AI usage, but one that may be overlooked is deliverability. If you aren't keeping an eye on your email deliverability, all your hard work can be quickly wasted.
   

   Solution: Include deliverability metrics in your AI optimization goals, and train your AI systems to flag potential spam triggers and balance engagement tactics with deliverability best practices.

6. Failing to test AI-generated content before deployment

   
   Risk: We're all about personalized content—the more personalized, the better. However, you want to make sure that your AI-generated content isn't going a little too far. Remember, it doesn't naturally have the same common knowledge as you, the marketer.
   

   Solution: Establish clear boundaries when it comes to personalization and ultra-personalized emails. Use AI to identify the level of personalization that drives engagement without triggering privacy concerns for different segments of your audience.

7. Creating a disjointed customer experience across email and other channels

   
   Risk: You might have a good thing going on between your emails and AI, but is it aligning with your other marketing channels? Differing tones of voice, types of content, and product focus across your marketing department leads to messy marketing, content chaos, and general bad (and confusing) vibes when it comes to customer experience.
   

   Solution: Use AI to help ensure consistent messaging across the board, offers, and tone across all customer touchpoints. Develop comprehensive prompts that reference your omnichannel strategy and customer journey maps.

   
   Check out our quick guide to using AI for content consistency.

AI for email marketing: What's next?

AI can be used in every marketing department, helping the whole team work more efficiently, productively, and creatively. Not only that, but AI can be used at every step of the marketing lifecycle too.

Check out full AI for marketing: Dos and don'ts content series (and see if you're making any of these mistakes).