Back to the Basics: How to Avoid Being Creepy with Your AI

The uncanny valley, deepfakes, hallucinations, scams. As exciting as the surge of AI has been recently, it hasn’t come without risk — and that risk hasn’t always come with reward. The everyday buyer remains wary of artificial intelligence’s role in the consumer experience, and one wrong move could be the difference between increased revenue and increased churn.

Marketers everywhere have spent the past few years thinking about how to leverage the power of first and third-party data to create more personalized experiences that build trust with their buyers. And much like finding the right balance between being personalized and being creepy with data, adopting AI requires a similar give and take.

Like any new brand, product, or service, the success of AI adoption hinges upon consumer value and consumer trust. And if consumers don’t trust your products or services, they won’t discover (or stick around to experience) their value. In fact, a recent study found that 84% of customers would consider leaving a vendor after a breach in digital trust. Conversely, research also shows that 46% of US consumers would pay more to buy from brands they trust. With AI playing an ever more prominent role in our digital lives, it’s imperative that vendors focus on building trust in their implementation of AI as well as building value with interesting solutions. When enabled with trust in mind, your use of AI and the resulting customer experiences will establish you as a leader in an incredibly dynamic space.

I’d like to offer four recommendations to guide you in your AI journey: Be Transparent, Be Human, Be Goal Oriented, and Be Authentic. (Not coincidentally, each of these recommendations can also make traditional marketing more successful.)

1. Be Transparent

The first component of avoiding a creepy AI is something a simple sentence can solve: Tell your buyers when they’re interacting with an AI solution.

If it’s an AI chatbot, for example, you can start the conversation with, “Hi, I’m [Company Name]’s AI bot. Anything I can help you find today?” Just this simple acknowledgement is enough to make a site visitor feel that they are in control of the conversation.

No one likes the wool pulled over their eyes, and by acknowledging you’re using an AI upfront, customers are more likely to understand, and be a bit more forgiving, should the AI hallucinate or otherwise say something awkward.

Furthermore, after introducing the AI, give your buyers the opportunity to opt out of the conversation if they’d like to. As stated at the beginning, many people are still wary of AI as a whole, so they shouldn’t be forced into something they don’t feel comfortable doing. If it doesn’t make sense within your AI solution to have an opt-out, be clear about what other channels or options are available for the customer to achieve their goal.

By including an option to opt out of the conversation or to route in a human, you acknowledge that not everyone is comfortable talking to a “robot,” and you offer alternatives that will still help you meet your customer’s needs. That’s what it takes to be customer-centric. Besides, whether you’re using an AI or not, any forced conversation is unlikely to convert.

2. Be Human

Yes, by nature, AI is not human. But that doesn’t mean humans shouldn’t be involved. Quite the opposite, actually.

Humans play a pivotal role in managing the fine line between an effectively personalized AI and creepy AI. First, most of the AI solutions marketers use today operate via machine learning, meaning that the AI improves over time through all of the interactions it has. This also means that, just like humans, the AI needs to be taught what constitutes a good response and what constitutes a bad one — and it needs to know when it has made a mistake.

To ensure your AI’s responses are up to your brand’s standards, you will need to incorporate a human-in-the-loop feedback process. This process can and should include both employees and customers. On the employee side, this assisted interaction, where an SDR or BDR is involved, may simply consist of reviewing a generative response prior to sending to the customer. On the customer side, this process could include asking the customer at the end of the engagement if the AI driven interaction was accurate and helpful.

Naturally, this does become more difficult at scale and when AI is interacting in unmoderated ways with the customer. Even so, it remains important to log every AI interaction and make it available for review and analysis. Training models is a never ending process and your models need your expertise to ensure proper training.

By stopping creepy before it happens, you lessen the risk for future negative situations sneaking through, and your customers will be much more likely to see the benefits of your AI solutions.

Marketing Technology News: MarTech Interview with Meagan White, Head of Marketing at Kibo

3. Be Goal Oriented

The success of any product implementation largely depends on how much thought went into why the implementation matters and what it’s meant to accomplish. That is to say, there needs to be intent behind the action.

This goal-oriented mindset is crucial to your success with AI. Don’t adopt AI just for the sake of being trendy. Think about where you see AI enhancing your customer experience or magnifying the ability of your employees to succeed. Take time to weigh the pros and cons, and don’t be afraid to ask vendors questions about their AI solutions. By taking such a thoughtful approach, you reduce the risk of careless mistakes that could threaten the trust you’ve already spent so much time working to build with your buyers.

When considering an AI solution for your business, start by first looking at your existing marketing processes and identifying areas for improvement. For example:

1. Are you struggling to understand your buyer’s journey? Look for intelligence tools that put your website under a microscope and reveal what pages convert the most and what topics buyer’s are seeking out. You can use this intelligence to inform your brand messaging, to provide more context to your sales team, and to predict what the next-best actions will be for your site visitors.
2. Are your customer-facing teams spending a lot of their day answering the same questions over and over again? Look for an AI product that you can train to answer these questions in your brand voice right on your website — and let your human teams get back to focusing on the bigger, more impactful, projects.

By keeping your overarching business goals top-of-mind, you’ll be able to think more strategically about what the right AI looks like for your business, and prevent breaches in trust along the way.

4. Be Authentic

Have you ever had an experience where someone you knew and trusted was less believable because they weren’t acting like themselves? Perhaps coming across as too enthusiastic or too  confident? You’re maybe caught off guard, unsure why the change occurred, or unsure how to react.

That’s a risk you run with your customers if you adopt uncustomizable AI.

In order to avoid being creepy with AI, it’s important that your marketers look for solutions that allow you to customize the AI’s messaging to fit your business’s tone of voice driven by your differentiated brand and content. Your buyers anticipate a certain style of response from you and, if the AI’s response does not match that, your authenticity comes into question. On the flip side, by maintaining consistent brand messaging across all channels, you build more trust with your buyers.

When you combine customized branding with human-in-the-loop moderation, you ensure that what the AI is saying is not only what you want it to say but how you want it said. Consistency generates trust, and making sure your AI is integrated into your overall brand experience plays a big part in that.

Human Relationships Reign Supreme

At the end of the day (and perhaps ironically) avoiding creepy AI means going back to what makes human relationships successful: being honest & transparent, achieving goals, and being authentic.

So, instead of looking at AI as a separate entity or add-on to your marketing processes, approach AI as a technology that enhances what’s already working for your team. If you’ve established a strong brand voice, are confident in your messaging, and like the way your buyer relationships are going, then look to AI as way to scale your successes, not replace them.

As long as you keep your customers at the center of your AI strategy, AI won’t be creepy — it will be your new best friend.

Marketing Technology News: Under The Hood Of The Adnami Attention Score