5 AI-Powered Breakthroughs That Will Help Sensory and Consumer Scientists Scale New Heights
This article is part of Aigora's "Original Content" series, which consists of our original thoughts at the intersection of consumer science and artificial intelligence!
I've been lucky. After I finished my Ph.D. in math, I wasn't sure what to do with my life. I was mainly sure I didn't want to be a pure mathematician having seen firsthand how awkward pure mathematicians can be (and knowing that awkwardness was already an issue for me). So I did what I usually do when I'm not sure what to do, which is to ask my father, Daniel Ennis, for advice. As luck would have it, my father had collaborated extensively with a psychology professor, Greg Ashby, at UC - Santa Barbara, where I had just completed my degree, and he suggested that I ask Greg for advice. By this time, Greg had transitioned into computational neuroscience and, as luck would have it, Greg was looking for a mathematician for his lab. Specifically, Greg had designed a neurobiologically-motivated model of human category learning, and he was interested in having someone program it. As luck would have it, I had prior Fortran programming experience so, lucky me, I spent 3 and half years working on neural networks for my postdoc, including this publication.
Following my postdoc, I worked for eleven years as a market research consultant for my father's company, The Institute for Perception, where I was again lucky as I was mentored by both him and Benoit Rousseau - under their guidance I was honored to receive the 2013 "Researcher of the Future" award from Food Quality and Preference. And now, in 2019, I'm lucky that I have the support of so many people, including my father and my wife, to launch this new company, Aigora, dedicated to helping sensory and consumer scientists prepare for the rise of artificial intelligence (AI).
I tell this story out of gratitude, yes, but also because it's relevant to today's topic, which is the main ways I see AI impacting sensory and consumer science in the near future. The work I did during my postdoc focused on human category learning, and hence was called psychology, but the same activities completed now would just as easily be called AI. So, once again, lucky me, I just happened to gain hands-on experience in the area that is now poised to transform society. It's perhaps out of this sense of how lucky I've been that I feel called to help others to succeed during the transition period our society is currently undergoing. Since it's sensory and consumer scientists whose problems I understand best, Aigora is my way of giving back.
For today's post, we continue to work our way up the "Pyramid of Preparation" we discussed last week, and now we find ourselves at the top. There are certainly applications of AI that are beyond our current conception, so today's post is not meant to be a glimpse into the far future. Instead, I have selected five breakthroughs that I see sensory and consumer scientists being able to enjoy either now or in the near future, thanks to AI, improved data collection, and other related technologies such as blockchain, robotics, and 3D printing. Of course, the actual deployment of these technologies will depend on many factors, but I hope that today's article raises awareness of the possibilities that AI offers sensory and consumer scientists. Let's begin!
Breakthrough 1: Closing the gap between qualitative and quantitative research
Consumer scientists have historically had to choose between collecting a large amount of low-resolution data or a smaller amount of high-resolution data. Tools for textual analysis and the possibility of conducting video interviews online following a survey undertaken previously have helped researchers collect qualitative information on a larger scale, but what has happened to-date pales in comparison to what's coming. Specifically, it's now possible to wrap chatbots inside Internet-based surveys, to ask follow-up questions in real time and as a function of respondent answers. Moreover, with advances in speech, facial, and even emotion recognition, it's straightforward to conceive of surveys being conducted directly on countertop AI-powered speakers, such as Amazon's Alexas, with conversational follow-up, and with emotional responses captured in real-time along with numerical and textual data. Samples could be distributed at a central location and evaluated in context with Alexa conducting the survey, recording the responses, and engaging in follow-up questioning. Once these technologies are even just a little further along, we're going to witness a sea-change in how we conduct consumer testing.
Breakthrough 2: Large-scale, low friction collection of behavioral and diary data
As costs decrease and the capability of tracking and processing usage data increases, sensory and consumer scientists will become free of the need to rely on the self-reported behavior of consumers. Does someone actually consume yogurt three times a week? Do they really clean their cat litter pan everyday? In the not too distant future, there will be large panels of consumers on whom data are continuously collected. These panels will be ideal for surveys as there will be little to no screening required - the respondent behavior will already be known. Consumer packaged goods (CPG) companies will also maintain private panels to record every aspect of their own brands' usage - in addition to optimizing brands around current usage behavior, learning non-standard product usage will be a source of innovation.
Breakthrough 3: Huge efficiency gains for quality control and quality assurance
Using vastly enriched datasets, sensory and consumer scientists will develop models to identify out-of-spec samples quickly and efficiently. Machine vision is already in use for inspecting samples visually, for example by Domino's, but as models for predicting sensory characteristics from instrumental data (such as Carlsberg seeks in their partnership with Microsoft) improve, quick instrumental measures will increasingly be used to categorize samples as acceptable or unacceptable. These same models will be used to make shelf-life projections as the data they are built upon grow in resolution. Also, consumer research can partner with these models in search of thresholds for consumer acceptability, potentially obviating the need for at least some discrimination testing and helping sensory keep up with the increasing speed of business.
Breakthrough 4: Supply chain revolution
Building on the use of models for quality assurance described above, the addition of advanced delivery systems and robotic farming will allow supply chains to become leaner while simultaneously improving the quality of the ingredients delivered. Moreover, the supply chain will be more clearly documented, allowing recalls to be less frequent and more targeted. The net result of this supply chain revolution to sensory and consumer scientists will be higher quality, less expensive, and more reliable ingredients with which to develop new products. And, finally, this revolution will also affect the consumer-facing distribution of products, allowing for targeted delivery of niche and freshly-prepared products.
Breakthrough 5: Targeted customer personalization
Recommender systems - such as those employed by Google, Amazon, and Netflix - are already one of the most widely-appearing forms of AI in society today. Even so, as the amount of data available increases exponentially and as consumers increasingly make decisions on configurable interfaces such as phone applications or menu screens, we're going to see we've only scratched the surface of our ability to recommend the ideal product to a would-be consumer as a function of everything we know about them in the decision moment, including the time, day of the week, weather, and perhaps even their mood (simulating service only typically available at boutique vendors such as this one). For example, Netflix customizes the very thumbnails its members see. Combining these custom recommendations with improved distribution systems - and potentially even the ability to 3D print products in the customer's home - we dramatically improve the value we can offer to our customers, winning both their loyalty and business.
Okay, that's it for today. I could say much more on each of these topics and will in future posts. Moreover, I haven't even discussed how virtual and augmented reality can be used to improve experimental validity or how natural language processing can be used to improve consumer education. Finally, there are other AI-powered breakthroughs for improving work processes that are outside of the scope of this article, but will also be covered in future posts. For now, I hope this list stimulates your thinking and opens your eyes to at least one possible breakthrough of AI that you hadn't previously considered. And, if the intersection of AI and consumer science is of interest to you, make sure you check out our "Eye on AI" series that we publish each Friday.
That's it for now. If you'd like to receive email updates from Aigora, including weekly video recaps of our blog activity, click on the button below to join our email list. Thanks for stopping by!