key

GEARING UP FOR ARTIFICIAL STUPIDITY

By: Shravya Kaparthi, Director, Analytics and Decision Sciences, RAPP Dallas

Artificial intelligence is revolutionizing consumer behavior in almost every aspect of the marketplace. The possibilities are exciting and virtually endless. Consider how the mass adoption of IoT smart devices has made it that much easier to restock the pantry, schedule appointments, and get simple product recommendations. But what about the times when AI gets it flat wrong?

Forgive the technology as it works out the kinks. It’s far from perfect in its current form, and it’ll take quite a while to iron out the wrinkles. Our impatience with AI can be telling: How many of us address voice assistants with pleasantries versus barking exasperated orders? How many of us lose our cool when Alexa does not understand what we are trying to ask and call it stupid? Many of us are cynical when an automated machine says in an amicable voice that it is “still learning and would appreciate some feedback.”

Idiotic machine! Right? As marketers, how do we work around this? It’s important to note that the lack of empathy goes both ways.

The Limitations of Artificial Intelligence

Let’s get hypothetical. From a customer experience point of view, AI interrupting the music to which a baby is napping with a programmatic “relevant” ad spot is extremely inappropriate and just cause for an outburst. At the same time, yelling back at the voice assistant is not exactly a productive response.

Nor is pulling our hair out when Netflix makes a terrible suggestion on the basis of previous consumption habits. We shouldn’t expect perfection from a nascent technology. News flash: Machines, like humans, make stupid mistakes.

As change agents, marketers have an obvious role in the development of artificial intelligence, encouraging certain behaviors based on identifiable gaps in the consumer’s life. But this role continues to evolve with each passing technological advancement. From consumer insights to performance management, technology now powers marketing.

While it’s important to keep pace with this AI evolution, you, as a marketer, are in the business of ideas. The onus is on you to get back to the basics and remember that the ultimate goal is solving problems for the consumer.

Solving problems, of course, requires empathy — that, and curiosity. You need to be curious about people beyond their categorical demographics. AI cannot — and will not ever be able to — do it alone. We humans need to try to understand consumer’s struggles, dreams, and beliefs. Each of these facets will make it easier to connect with them and inspire that next big purchase.

Using AI to Improve the Customer Experience

To crack the code of empathy and overcome AI limitations, you need to “train” artificial intelligence with emotional intelligence in order to further enhance human intelligence. Because the fact of the matter is, artificial intelligence doesn’t need to be any more intelligent than it is today. It just needs to care.

And with the rapid pace at which AI is growing, sometimes the technology is being tested in the wild before it’s properly vetted, which can lead to myriad problems: inherent biases, privacy infringements, indirect safety threats, and more. Even carefully crafted AI systems can act in ways the developer never anticipated.

These systems work on reinforced learning, and there is a version of AI that takes what you do and amplifies it. As such, your mistakes will also be amplified. So it’s more important than ever to collaborate with others and think about what to build rather than how to improve what’s already available.

Instead of incremental benefits like driving efficiency or mechanizing routines, focus attention on how AI can be used to help solve problems, build competitive advantages, and transform the customer experience.

Keeping the Human Element Intact

Let’s say a telecom provider is looking to improve customer service. Using AI on its own won’t solve the problem without some level of human touch along the entire value chain. Machines can only enhance the experience — and even then, you must first have a strong process in place for machines to provide the expected outcomes.

That’s why the savviest option will be to weave AI into the fabric of your approach. It ensures that the customer experience isn’t just a one-size-fits-all endeavor. You need that human to make the connection. Otherwise, you end up with logically correct artificial stupidity rather than empathetic human rationale.

It’s all about overcoming the limitations of artificial intelligence. The following strategies should help you circumvent the overreliance on AI:

  1. Lead with people. Human-centric design is a more creative approach to problem-solving, where the design process starts with specific people in mind. That way, the product or service becomes an almost tailor-made solution, completely fitting the user’s unique needs. Learn as much as you possibly can about the people you’re designing for and use this information to identify opportunities for design solutions.
  2. Think bigger. The customer journey is a complex one, sometimes making it difficult to understand how the target customer arrives at a purchase decision. Use AI to provide consumer context and develop profiles to better reveal the motivations and behaviors of a target audience. The ultimate goal is to discover human truth, not just short-term solutions for improving processes and efficiencies.
  3. Develop backup options. People don’t build relationships with algorithms. But they do with other people, which is why human intervention is still important. Take customer service, for example. No matter how advanced, AI can only answer so many questions. A point will come when an actual person must step in. Design a seamless human-machine integration across all customer touchpoints to ensure you’re using AI to improve customer experience.
  4. Underpromise, overdeliver. Building AI into the organizational culture is always a work in progress. After all, the limitations of AI technologies can lead to outcomes that don’t necessarily align with the intended goals. Never forget the drawbacks of artificial intelligence. The cost alone can be restrictive. So account for any risks involved with new AI initiatives by managing the expectations of all stakeholders.
  5. Embrace failure. After all your efforts, machines can still disappoint — just like humans do. Training AI for customer engagement is an ongoing process with multiple iterations. Patience is key. Accept the mistakes made by computers and leave room for periodic failures. AI can only perform what’s been programmed. Without the intuitive abilities of a human, it can’t judge whether an action is right or wrong. It simply acts.

Technology will keep changing, incrementally improving each year. But for all the iterative benefits AI can provide the marketing industry, it will always fall short of being human. Sure, it can capture human insights and can even make sense of learnings. From there, however, machines should pass the baton. You don’t want AI doing things that humans can do better — and vice versa.