Can AI Help in Finding Optimal Pricing?

October 20th, 2020
Rick Quinlan | Vice President
Hero Image: Can AI Help in Finding Optimal Pricing?

Technology is simply brilliant! If I didn’t already embrace that fact, “60 Minutes” reinforced that upon me with their article about Kai-Fu Lee, the “Oracle of Artificial Intelligence” recently.

A company of Mr. Lee’s, Face ++, has deep-learning facial response programs that can tell educators which students are engaged, bored or confused during classroom lectures. Teachers can see at what point during lecture these responses happened, and can follow up with the individuals.

And, modern kitchens can re-order supplies for you, such as your fridge noting the milk or eggs are empty, or when you have used any of the food you purchased. The fridge can automatically contact your supplier and new food will arrive before you even knew it was needed.

But, the practical success of technology has its limitations. The teachers can be alerted as to which students were excited/confused by the lecture, but the software does not know “why” it happened, and doesn’t know anything more than the facial response identified. “60 Minutes” notes that “a typical AI system can do one thing well, but can’t adapt what it knows to any other task.”

The utility of augmented commerce such as the automatic re-ordering of supplies does only that, re-order without taking in a human element and the idea of automatic replenishment is sensible. But, will it succeed in the market?

What if you knew a recent survey by job finder CareerBuilder identified that 78% of American households are living paycheck-to-paycheck. So, beyond basic bills like housing, utilities and communication, the “automatic” replenishing of supplies just because they are no longer available may not be a usable enhancement. Not to mention the family might want to have some variety in their meals compared to the previous week.

AI will bring advantages to the world of consumer insights as well. However, AI is the culmination of deep learning from existing data. TRC uses predictive modeling but brings more insight combining such findings with consumer opinion.

I don’t see AI designing conjoint studies to assist in product development or finding optimal pricing. Nor will it decide that Max-Diff would be the best method for solving a particular business challenge, but it isn’t hard to see where AI could help. For example, if webcams can gather the sort of facial expression data that Mr. Lee’s system does, we could use it to understand how engaged respondents are. If this were combined with data on actual consumer experiences, perhaps we then turn complementary data points into consumer insight and business direction.

Technology is brilliant, but can only solve the problem it addresses. So, in order to have AI succeed, people insights are needed. Combining what we know about people with what can be done technically seems necessary for ultimate success.

Mr. Lee told “60 Minutes”, “I believe there is a lot of things about us that we don’t understand. I believe there’s a lot of love and compassion that is not explainable in terms of neural networks and computation algorithms.” Even the “oracle” of AI believes we need more information.