The emerging company was also chosen as one of the 10 startups for AB InBev and Techstars Connection accelerator program, which will provide the early-stage businesses with access to mentors and potential investors.
Analytical Flavor Systems set out to make flavor a predictive element of the manufacturing process, something that current methods of sensory science has not been able to achieve, according to founder and CEO Jason Cohen.
“We believe sensory science as it’s practiced today is failing companies,” Cohen told BeverageDaily. “We’ve made flavor a predictive measurement.”
The company has active clients in the beer, coffee, and soon chocolate industries, with sights set on wine, spirits, soda, and Asian alcohol.
Treating flavor as predictive science
Cohen identified a gap in ability of traditional manufacturing processes to provide predictive flavouring of consumer packaged goods. “Food and beverage CPGs produce inconsistent products largely because they don’t have as much control over the process as they’d like,” he said.
“This is particularly evident in producers’ inability to manage the effects of raw ingredient deviations and degradation, as well as flaws that occur throughout the manufacturing process that adversely affect the product’s flavor-profile. This lack of control can significantly hurt the brand and product sales because consumers demand a consistent experience every time they buy that product.”
After about three years of research, Cohen and his team at Analytical Flavor Systems figured out a way to translate a human’s perception of what he or she is tasting into a numerical system.
Using artificial intelligence (AI), the company modeled flavor profiles and the underlying chemical through a machine learning process control solution to target higher taste consistency.
“It’s easy to dismiss flavor as subjective, but we know that that’s not true because there’s international successful products whose brand relies on flavor consistency,” Cohen said. “The flavor profile is the key selling proposition of any beverage brand.”
Preventing flavor drift
The technology the company developed, called Gastrograph, provides predictions for flavor profile formation and optimization, allowing early detection and correction of batch-to-batch deviations, flaws, taints, and contaminations at any stage of the production process.
The intelligent software learns to control the manufacturing process, updating times, temperature, and ingredient additions automatically – ensuring the finished product hits the target flavor profile every time.
This type of technology is incredibly helpful to large beverage companies, especially brewers, which have multiple production facilities scattered across the US.
Large brewers are susceptible to “flavor drift” in which the aroma and flavor composition can deviate over time because of the slightly different manufacturing practices done at each facility, albeit unknowingly, Cohen said.
“If you have five or six different breweries, that’s five to six different versions of that beer. You’re degrading brand value” because consumers today demand an identical flavor profile no matter where they are drinking the product, Cohen said.
“The AI is just sorting out what humans are saying they’re tasting,” Cohen said. “We just created a way to understand to human sensory perception.”
Expecting accelerated growth
Analytical Flavor Systems expects to benefit greatly from the 17-week-long accelerator program.
“We’re growing pretty quickly. We’re signing on a bunch of much, much larger breweries, coffee roasters, and chocolate manufacturers, and we’re looking to expand into spirits and wine,” Cohen said.
“We expect that we’ll be in a number of Fortune 500 food and beverage companies.”