Dear [insert name here],
Long time, no see! I hope you’ve been doing well! How is the family? What’s your dog’s name again?
OK, let’s start over. I’m a real person, not an AI agent…but I guess you’ll have to take my word for that. And I’m sorry that I don’t know you, or your family, or your back story…but I’m also not going to pretend that I do.
You don’t know me!
I like to think I can spot an AI-generated marketing email from ten paces. I usually send them straight to my junk folder before reading past the subject line, (unless an AI spam filter beats me to it), but a recent one caught my attention.
This story starts with a chatty email I received from someone I’d never heard of.
He called me by name, mentioned the city I grew up in, dropped a comment about a recent company milestone, and finished by asking me if a particular local restaurant was still open, having heard great things about one of their dishes.
Surely this was from someone who knew me, or at least had taken the time to find out about me, and wanted to show they were genuinely interested in me? Call me a cynic, but to me it seemed more like an AI bot that knew how to trawl my LinkedIn profile and company website, and wrap their marketing spin in platitudes.
As it transpired, it was from a company offering “hyper-personalized”, AI-powered email marketing campaigns. Their website promoted such features as, wait for it…including personalized comments, company milestones, and a restaurant recommendation, including its top dish, “to start a conversation”.
Out of interest, I searched for information on warning signs that an email has been generated by AI. Depressingly, the first link I clicked on was to an article that presented as objective research, but was actually spruiking yet another AI-generated emailing product. They claimed they could make theirs sound less AI-generated…by adding more personalization! Their claim:
Email marketing is evolving - Machine-learning tools are unlocking hyper-granular personalization!
Is that really something to celebrate?
And Mahesh doesn’t own me!
The thing that first caught my attention in the original email was a reference, in the subject line no less, to “Mahesh, your owner.” This intrigued me:
I don’t know anyone called Mahesh, but I do know our company owners. If they’d recently sold out to a Mahesh, I’m pretty sure they would have mentioned it to me.
So how did they get this so wrong, and who is Mahesh? I did a simple Google search. Who does the web think owns our company?
In no time, I was on a site providing a company profile. Most of the content looked about right, clearly lifted from our website or other public domain sources. The head-scratcher was a list of staff members and executives. Most were my colleagues, but the owner, CEO, and some others were people I’d never heard of, and we are a small company. Huh?
A bit more searching, and I learn that these names, including the mysterious Mahesh, were real people, associated with companies on different continents that have no relationship with us, but share almost identical business names.
My conclusion:
- My marketing email shows all the signs of having been created by an AI agent.
- That agent ‘thought’ it could sound more authentic by trawling the web for supposedly relevant ‘personal’ content to include.
- That content happened to have been created by yet another AI agent…which had itself been trawling the web for supposedly relevant content.
- Content that started with a grain of truth, but was then misinterpreted and converted into nonsense.
What a slippery slope. A descending spiral into a highly flawed feedback loop, siphoning up information of ever-diminishing credibility, and regurgitating it as fact.
The Moral of the Story
It's clear that AI is going to be an increasingly important part of business. The vote is in on that, and there is no turning back. There is no denying it has a seductive potential to increase efficiency and productivity. But a warning:
If we surrender to the appeal of this shiny new toy to the extent that we lose our ethical compass, we will all pay a price.
Trust and respect are the bedrock of successful relationships, whether in our personal or business lives. These can take an age to build, but can be destroyed in an instance, or with an email.
Is a pseudo-personal, AI-generated marketing email that pretends to be something it’s not any more ethical than a deep fake video designed to deceive and manipulate?
The example in my case study above was so inept and obvious as to be almost laughable, but the more convincing these become, the more sinister their potential.
My message to marketers, and their AI bots, if thinking of embarking on a “hyper-personalized” email campaign:
Sorry, but I don’t want you to 'treat me like a real person’ when it’s patently obvious that you yourself aren’t one.
My Verdict:
Hyper-personalized AI marketing isn’t too good to be true. It’s just not true enough to be good.
