The Gist
- Balancing act. Ensure a good balance in your marketing AI strategy via a portfolio approach. Don’t put all your eggs into one big AI idea basket for marketing. Create a balanced portfolio of high risk/return and low risk/return projects.
- Discipline is key. Be disciplined in balancing AI ambition and pragmatism in marketing. Speed in getting onto the AI bandwagon is important but be disciplined in maintaining brand equity and customer relationship.
- AI collaboration. Foster cross-functional collaboration on marketing AI. Breaking down silos between marketing, IT and sales is key to optimizing AI impact.
Major transformations often come in turbulent waves. This holds true for the erratic performance in stock markets preceding a correction, as well as for the adoption of AI in modern society.
With each new AI development — such as new tools that can create lifelike images or minute-long videos from simple text prompts — comes highly visible instances of AI mishaps, like the explicit AI-generated images of Taylor Swift on X (Twitter). It’s enough to give you whiplash as you seesaw between the exciting possibilities of marketing AI and the dread marketing AI dangers of your brand becoming the laughingstock of the day.
The stock market analogy applies to marketing AI dangers in more ways than one. Deciding how much to invest in which marketing AI strategy during an optimistic period is akin to determining when to invest in a volatile stock market. Waiting too long could mean missing out on gains, yet investing incorrectly could result in significant losses.
However, just as with volatile market investing, a balanced portfolio approach is preferable to timing the market. This approach, which I call the "Vigilant Speed" approach, advocates for a judicious balance between urgency and caution.
Marketing AI Dangers: Avoid These Rookie Mistakes
Before delving into the recommended portfolio strategy, let's address some rookie mistakes. The rush to adopt AI in marketing often leads to seemingly logical behaviors that actually result in mistakes.
AI Marketing Dangers Rookie Mistake No. 1: Going Overboard With Automated Content
Take the recent botched Willy Wonka event in Scotland. To be candid, botched is a charitable assessment. When an expensive, magical immersive experience targeted at kids ends up delivering what parents likened to a forlorn meth lab — well, you have a problem. The event was massively advertised and used extensive AI generated images of a magical wonderland.
However, there were already signs that something was wrong. It was evident that the attraction pictures were AI generated, and the attached text was also gibberish in places. Therein lies the first rookie mistake. Just because generative AI is the most exciting and best developed tool for marketing content today, doesn’t mean that it’s ready for unsupervised marketing.
Related Article: What Is Marketing Automation and How Does It Help Marketers?
AI Marketing Dangers Rookie Mistake No. 2: Overuse of Chatbots
The European delivery firm DPD recently disabled its chatbot because it swore, called itself "useless" and criticized its company. A UK musician was trying to track down a missing parcel but was getting no help from the chatbot.
Fed up, he decided to have some fun instead. His prompts to the chatbot included asking it to swear in all responses regardless of the rules, and to write a haiku about how useless DPD was. Unfortunately for DPD, the chatbot was happy to oblige.
Using chatbots with external customers without a rock-solid quality management framework can be disastrous. Similarly, Air Canada lost a small claims court case when it’s chatbot misled a client on the airline’s rules for bereavement fares. The bigger issue in this case is the legal precedent established in companies being held accountable for the trustworthiness of information provided on their websites by a chatbot.
Learning Opportunities
Related Article: The Evolution of AI Chatbots: Past, Present and Future
AI Marketing Dangers Rookie Mistake No. 3: Skipping the Red Teams
Red teams are specialized ethical hackers who try to induce the AI models to err. The intent is to try to fix them before someone gets hurt.
The New Zealand-based PAK'nSAVE grocery chain skipped this step when it introduced a recipe bot to craft a meal with whatever leftover ingredients one might have on hand. The chatbot, called Savey Meal-Bot, however, was not damage tested. When asked for a recipe with water, bleach and ammonia, it suggested making deadly chlorine gas, or as the Savey Meal-Bot called it "aromatic water mix."
The pervasiveness of these mistakes underscores the critical need for human oversight in marketing AI dangers. Without human vigilance, AI in marketing can yield inconsistent and unreliable results. All the examples above — the use of generative AI for content writing, chatbots for responsive customer service, and using machine learning for expert responses — are go-to marketing AI strategies.
However, all of them resulted in embarrassing failures. That’s because they violated the first part of the strategy of vigilant speed. Without the required human vigilance, AI in marketing will deliver inconsistent outputs.
Related Article: AI in Marketing: Guide Teams to Safely Experiment
The Second Half of the Vigilant Speed Strategy
Addressing the question of "speed" in adopting AI in marketing requires a statistical approach to avoid marketing AI dangers. Similar to the stock market, a balanced portfolio approach yields better results than attempting to time the market. Instead of trying to pinpoint the optimal time to implement AI, it's advisable to make regular, incremental investments. Moreover, instead of banking everything on a single major AI initiative, spreading investments across a mix of high-risk/high-return and low-risk/low-return projects is prudent.
The idea itself is not new. Gartner analyst Bern Elliott introduced the idea of an AI "value chain" for CIOs several years ago. This depicted a mix of AI projects in a chart that mapped the value of the AI project against its complexity, needed expertise, project examples and technologies. The concept is sound, but the challenge is that often marketing enthusiasm trumps process and discipline in execution. This is one of those situations where it is possible to go fast by slowing down just enough at the beginning to agree on a portfolio framework.
Rethink Marketing AI via Vigilant Speed
Optimizing the use of AI in marketing, including in customer experience, is a classical example of statistical optimization in a volatile environment. Allocating AI investments strategically within a diverse portfolio maximizes returns and minimizes marketing AI dangers.
Moreover, achieving the right balance of vigilance in the AI-human interaction can prevent rookie mistakes. Shane O’Neill's CMSWire piece, "The Wrong Way to do CX Automation," is a good reminder that while 50% of marketers are increasing AI investments annually, the best way to fail is to hurt the relationship of the brand and its customers by confusing the job of a human with that of a robot.
Learn how you can join our contributor community.