A lot of AI talk in the email industry focuses on how it will help marketers create better content faster than ever before, analyze results and recommend tests, and generally make the email marketing process quicker and easier. All true, but the AI impact on email isn’t all positive for mailers. AI is also making email marketing more complex and challenging for mailers. Here are a few places where AI is changing the playing field for email marketers in unexpected ways.
AI Summaries
Email recipients are familiar with preview text. Many ISPs allow for a line of text to appear in the inbox along with the sender name and subject line. This preview text is determined by the sender. So, as a mailer, you get to decide what text to include in this preview section.
Preview text has its pros and cons for mailers. If the recipient finds the preview text interesting, they might be more likely to open the email. If it isn’t interesting, they may be less likely to open the email. But preview text is really just that – a preview. So, in general, it should help recipients identify interesting emails and then open them for further review. That’s likely a net positive for mailers.
Enter AI summaries. Various ISPs and third-party email apps leverage AI to review the contents of an email when it hits the inbox and provide a summary to the recipient before opening. Sounds a bit like preview text on the surface, but it’s a very different animal.
Preview text is a short introductory phrase or sentence to entice recipients to open the email, written by the mailer. It isn’t a full summary. It’s more of an augmentation to the subject line. AI summaries are full summaries of the email and, being written by an AI rather than the mailer, it adds some challenging variables into the mix.
AI may or may not do a good job of summarizing an email’s content. Some of these tools have shown that they don’t always deliver an accurate summary. Elements like humor can confuse them. Multiple CTAs or a complicated call-to-action may also create issues. Even a straightforward CTA may not come through properly in the summary. If the AI delivers a poorly worded or uninteresting summary, it could depress opens and further engagement from recipients.
A recent study from Validity sheds some interesting light on the impact. Setting aside whether a summary is good or bad, recipients may read it and then decide not to open the email. This drops open rates (a flawed metric, but still directionally useful) and may also depress downstream engagement measures. An engaging AI summary could mean a recipient simply goes off to visit the brand’s website without ever engaging with the email, making attribution for any action they ultimately take difficult to connect back to the email. In either scenario, the email marketer is being negatively impacted by the AI summary.
AI Assistants
For some email recipients, AI is doing a lot more than simply providing content summaries. A growing number of people are using AI as a type of virtual personal assistant. Part of that role may involve managing the person’s email inbox to varying degrees. These types of AI agents can evaluate the contents of an inbox, highlight messages that it determines need human review, and deprioritize or even delete messages that don’t meet certain criteria.
Yes, ISPs have been filtering emails at scale for years. It’s why so many messages end up in our spam folders, and others end up in promotion or social tabs, for example. The ISP is using tools (likely AI these days) to decide what to do with every email sent to a recipient’s inbox. To at least some extent, the rules behind those decisions are understood by email marketers, and they are generally consistent from one recipient’s inbox to the next on the same ISP.
AI assistants that review, highlight, and potentially delete emails from the inbox may do so based on specific guidelines provided by the human inbox owner or from the AI’s own learning about that person’s individual behavior. This is a new and more challenging type of message filtering. Each AI assistant may approach its role differently (in fact, they probably will), based on how it was instructed by the user. So, a message that reaches one recipient’s inbox may fail to reach another, based on individualized AI filtering.
AI Buyers
This is still in the very early stages, but a small percentage of consumers aren’t just using AI as a virtual assistant. They are having AI make purchase decisions and then carry out those transactions. On the surface, this sounds like a positive for brands and companies whose products or services become part of an AI buying program. However, it may spell trouble for affiliate marketers in the email channel.
Imagine a situation where an AI assistant has been augmented with the ability to make purchases. An affiliate email hits the inbox with a great offer on a product the AI has been authorized to buy if the price drops below a certain point or the % off offer reaches a set level. The AI may simply trigger a visit to the brand’s website and make the purchase. This would likely bypass the email’s tracking link, meaning the sale wouldn’t be attributed back to the affiliate or even the email channel as a whole.
This practice hasn’t reached significant popularity yet, but it seems to be just a matter of time as people take advantage of AI capabilities to simplify their lives.
What’s Next?
AI is still really in its infancy, and it’s already disrupting industries on a daily basis. Marketers need to balance taking advantage of the ways AI can positively impact their businesses while closely monitoring and adapting to developments that challenge existing revenue models and processes. Those who take the time to watch what’s on the horizon will be best positioned to weather the storm or capitalize on new opportunities.