Using IBM Research machine learning algorithms for product composition, McCormick said it is “ushering in a new era of flavour innovation” that has the potential to change the course of the food industry.
Speaking to FoodNavigator, Dr. Faridi said that the new AI platform brings together all of McCormick’s proprietary sensory science and taste data, which includes decades of past product formulas and millions of data points related to consumer taste preferences and palettes.
“We have more than 40 years of documented data, sensory and taste data, in our system. We were one of the early adopters of sensory methodology… This data was always in silos, it was never shared. We also have very strong roots in chef and culinary [development]. We wanted all these things to come together in the pursuit of ‘what is next in flavour?’”
The new AI-enabled product platform is a learning system that combines the experiences of McCormick’s global development workforce, which today is made up of 500 people, at 20 labs, in 14 countries across five continents.
“We wanted to create a learning system that can learn from the experiences of everybody around the world and give feedback to everybody around the world. There is no way we could do this… without a learning machine that can look at the data, give feedback, learn from it, improve it and come back with a new suggestion,” Dr. Faridi revealed.
IBM was able to take McCormick’s wealth of data and build algorithms that can generate new recipe formulation suggestions. “Our team has created new algorithms so this really is innovation in AI that is being applied to this pioneering work that McCormick is doing with us,” Dr. Robin Lougee, research scientist at IBM Research, explained.
“The richness of their data is a fundamental aspect of the success of this project. People may have thought about it but it was really the commitment and rigor that McCormick has in terms of capturing and recording that data that enables us to even be able to think about doing this kind of AI.”
McCormick first developed a vision to create “one global lab” to “create the R&D of the 21st century” a decade ago. The project has taken four years to construct. “We think by 2021 every member of R&D in all 14 countries will be connected to this machine learning system,” Dr. Faridi revealed.
Better, faster product development
The AI programme takes “minutes” to generate “dozens” of formulation suggestions, Dr. Lougee said.
For McCormick, this will streamline the product development process – opening the door for it to bring more products to market at a faster pace.
Over the past 18 months, McCormick and IBM have developed the system enough so that the product development team can test it. Dr. Faridi said that it has already proven it can deliver efficiency gains.
He explained: “A developer always starts with a base formula and builds on it. Then sensory will test it and come back with suggestions. Somewhere between 50 and 150 iterations is the norm. We have found that with this we can cut that process by two-thirds. We can make it much faster. I think two-thirds is conservative: four or five years down the road will be much better.”
And it is not just about efficiency, it is also about creativity, IBM’s Dr. Lougee told this publication. “Our team in particular is interested in exploring creativity. Can computers help humans be more creative? Food is one of those areas where humans have always been exploring creativity.”
She elaborated: “Product developers sometimes have biases that the machine does not. We want to be able to help the product developers explore their set ideas to come up with the next iconic products. Thanks to the wealth of data that McCormick has, the computer can look across a pool of data that the human would be unable to read and reason over. The AI system can find things product developers may not have thought of…. It can take you outside your biases and comfort zones to look at things you might not have realised.”
Dr. Faridi concurs that the AI system is able to push the boundaries of product development in a way that human developers are not. “When we develop any product we have a hedonic target of 6.5 or 7. So the developer, when they get the sensory evaluation to 7, accepts that is ready to go to full commercialisation. The machine has no boundary on this. It is always looking for making it better and making suggestions.”
‘This is a paradigm shift’
Dr. Faridi believes that this will enable the food sector to respond to high levels of product churn. “One of the challenges that I have is how can we make a product that will have staying power in the market, how can we make it stick?
“This machine, because it doesn’t stop creating, gives options and novel solutions that are pushing the boundaries... It is a learning machine. Every day is better than the previous day. It is going to help us develop more preferred flavours, more icons, for ourselves and our customers.”
It will also allow McCormick to be more agile in flavours development, shortening the product pipeline and increasing speed to market as the group responds to the fast-pace of consumer trends. While today is the first time McCormick has shared details on the development, Dr. Faridi said that his contacts with food industry customers who have heard of the development at a “conceptual” level are excited by the possibilities it opens.
“It will help us for collaboration in our b2b flavour solutions business, tremendous collaboration with our customer base.
“This is a new way of thinking, a paradigm shift. Any time I talk to my customers, everybody is excited. But it is not a question of excitement. The best thing to do is develop products that win in the market and win against other companies. That is the bottom line and we think it is easily achievable.”
Leap-frogging the competition
McCormick has used the AI system to develop three products that it will be bringing to market under its consumer-facing spice brand. These initial one-dish Recipe Mix flavors include Tuscan Chicken, Bourbon Pork Tenderloin and New Orleans Sausage.
However, the opportunity goes way beyond this. “There are many possibilities. It is almost unlimited. I am always looking not to compete with someone. I want to leap-frog them, I want to create something that people cannot compete with. This will enable us to do this,” Dr. Faridis said, adding that he is unaware of any peer group companies who are yet close to developing similar technological capabilities.
IBM and McCormick intend to continue collaborating to extend the functionality of the AI system and increase the data it is built on to include the chemical composition of flavours.
“This isn’t done by any means, there are a lot of technical challenges that our team is working on and we are really excited to have access to the data and expertise at McCormick. We are looking forward to incorporating even more data and tools in the coming year and beyond,” Dr. Lougee said.
“The innovative technologies that IBM has developed and our knowledge of flavour is a marriage made in heaven. What we are sharing with you is really just the tip of the iceberg,” Dr. Faridis added.
FoodNavigator is hosting a webinar on the disruptive potential of AI and machine learning on 14 February.
Register to join us and our expert panel by clicking here.