AI in Marketing: Then and Now
Robots are literally taking over the world. Not actually – although today’s emphasis on efficiency indicates that we probably aren’t too far from the robot apocalypse. For the time being, advancements in Artificial Intelligence have been sweeping industries, from cars that can park themselves to email lists that segment themselves and send emails accordingly. But where did AI first find it’s marketing niche?
AI Back in the Day
AI earned a spot in the marketing world with behavioral targeting, helping online advertisers and marketers target audiences with specific traits, behaviors and characteristics. Grouping these human behaviors and tracking patterns between people possessing similar traits can allow AI to predict future behaviors. For example, the “you may also like” section of e-commerce sites suggests products that you might like based on shopping patterns of other users.
More recently, Yahoo’s automated insights have produced AI-generated content, tailored to users based on interests. Taking that behavioral targeting a step further, Yahoo’s Automated Insights Wordsmith platform uses natural language to generate user-specific sports content – pulling stats and draft summaries “at a speed and scale that would be impossible for human writers.” Content generators coupled with machine-learning technology make these pieces sound natural and organic instead of robot generated.
AI and Machine Learning
The above Tweet referencing an AI powered pizza robot generated some backlash from the tech community. “Look at this AI powered coffee robot,” the Tweet said. The following photo of a Mr. Coffee pot illustrated the point that Artificial Intelligence does not simply mean a machine. If so, AI has existed for as long as the machine-powered assembly line. We don’t call the ice maker a robot, why would we call another machine a robot?
In marketing, there’s a difference between AI and machine-learning. AI is the idea that a machine can be programmed to execute a task, and machine learning is the idea that programs can learn and adapt based on usage. For example, Google’s BERT update, allowing the search engine to better understand the nuances of human language, is a machine learning update, while Google assistant is powered by AI, but uses machine learning technology to improve task execution. Advancements in machine-learning only mean advancements in AI.
As a marketer, I’ve seen an increase in AI-powered programs in the email marketing space specifically. You’ve already seen how AI has determined the best send times – following an email’s journey from delivery to open to click. Send in Blue, for example, collects this data over time and eventually allows marketers to utilize this feature to optimize opens. Acoustic Campaign (formerly IBM Watson Campaign Automation) has this same function – STO or send time optimization.
In addition to AI’s ability to optimize send times, AI can help write subject lines. Using metrics from past emails, AI can predict which subject lines will perform best. For example, AI can assess the value in sending an email with a long subject line verses a short subject line. This could eliminate the need to run A/B tests and save resources. AI can even help craft the body of the email using pre-written copy, relevant populated (stock) images and links to sources.
AI Moving Forward
AI-written content has already existed, but I’m anticipating an expanded presence on email marketing platforms. Having a stronger presence in the sending optimization and creation of email campaigns will allow businesses with limited resources to better control their marketing. AI will enable marketers to handle larger volumes of email content and cater to more needs without allocating resources to staffing and time.
Instead of AI being solely a technological perk, like software, there will be entire businesses structured around AI. While the AI business would mimic the software business, there’s promise in AI businesses focusing on product (the AI itself) and service, since the technological realm is shifting towards a more automated approach. For example, unlike email marketing services like Acoustic Campaign that offer AI services in the platform, businesses will start selling AI. Once people analyze how to maximize profit margins, AI will take the business world.
However, the true downside of AI is it’s human replacement. Circling back to Yahoo’s AI-aggregated content, it’s apparent that the low price, quick turnaround and impressive content volume could easily replace real humans. AI does not need to collect a salary, and AI can scan and skim resources faster than humans. Will it eventually replace the need for human content creators and marketers? Only time will tell.