5 min
[caption id="attachment_38075" align="alignnone" width="940"]
Source:NatGeo[/caption]
Most of us are familiar with the packaging videos thanks to NatGeo’s mega-factories. The food and beverages industry giants like Pepsico bring over $ 60 billion in revenue per year. To meet the high consumption rate and to maintain the consistency in taste and texture of the product, companies go to great lengths by deploying state-of-the-art tech.
Machine learning has been put to use for improving user experience in products like homepods, mobile phones etc. But when it comes to food, there isn’t much room for experimentation, especially for legacy products like pepsi-cola or potato chips. Customers will find out any change in the experience. This unlike other products is more tangible and personal.
So, to maintain the consistency, at Pepsico, AI is being used to check the texture of the chips without destroying them. In an interview to a leading online portal on automation technology, Shahmeer Mirza from Pepsico R&D, gave a brief overview of the company deploys AI and ML.
Ever since its merger with Frito-Lay, Pepsico has more or less dominated the food and beverages industry. It has grown into a powerhouse that includes brands like Tropicana, Lay’s , Cheetos and Gatorade among many others.
To check the texture of chip in a non-destructive way, a photo acoustic apparatus was built. This apparatus includes a laser generating tool, an acoustic capturing device and data processing unit.
A laser is directed towards the chip which when struck produces disturbance in the immediate surroundings and creates pressure waves(sound waves).
These waves produce an acoustic signal which is captured and forwarded to the processing unit. Here, the acoustic signal is filtered and quantitative acoustic model is generated. This model consists of the key properties of the chip like how hard or fragile it is. A lot of data is generated during this process and then comes the challenge like making insights out of this data where machine learning models come into play.
Many other sophisticated tasks are being augmented with machine learning. For instance, testing the weight of a potato before and after peeling. For this the R&D team is deploying deep learning models with classifiers such as Random Forests. These neural networks are fed with the images of the potatoes and are trained to give an approximation of the weight by identifying the percentage of peel. This eliminates any chance of extra peeling of the potatoes.
AI has found its way into almost all departments at Pepsico; from maintaining tastes to avoiding wastage. It has given control over what goes into the process and how good the product is. Not only that Pepsico has one foot in to the managing the output as well, with its Snackbot. Snackbot is a robot that delivers chips on a university campus. This driverless bot manages to steer through different obstacles by employing machine vision techniques.
It is being churned out for optimisation, no matter how small and, in domains never before imagined. The transition to data driven strategies is rapid. From product development to sales and marketing, there has been a tremendous improvement and companies are filling their pockets or at least saving them from spilling.
Check out other ways in which AI is put to use here.
Learn more about the photo-acoustic apparatus here.
Source:NatGeo[/caption]
Most of us are familiar with the packaging videos thanks to NatGeo’s mega-factories. The food and beverages industry giants like Pepsico bring over $ 60 billion in revenue per year. To meet the high consumption rate and to maintain the consistency in taste and texture of the product, companies go to great lengths by deploying state-of-the-art tech.
Machine learning has been put to use for improving user experience in products like homepods, mobile phones etc. But when it comes to food, there isn’t much room for experimentation, especially for legacy products like pepsi-cola or potato chips. Customers will find out any change in the experience. This unlike other products is more tangible and personal.
So, to maintain the consistency, at Pepsico, AI is being used to check the texture of the chips without destroying them. In an interview to a leading online portal on automation technology, Shahmeer Mirza from Pepsico R&D, gave a brief overview of the company deploys AI and ML.
Ever since its merger with Frito-Lay, Pepsico has more or less dominated the food and beverages industry. It has grown into a powerhouse that includes brands like Tropicana, Lay’s , Cheetos and Gatorade among many others.
To check the texture of chip in a non-destructive way, a photo acoustic apparatus was built. This apparatus includes a laser generating tool, an acoustic capturing device and data processing unit.
A laser is directed towards the chip which when struck produces disturbance in the immediate surroundings and creates pressure waves(sound waves).
These waves produce an acoustic signal which is captured and forwarded to the processing unit. Here, the acoustic signal is filtered and quantitative acoustic model is generated. This model consists of the key properties of the chip like how hard or fragile it is. A lot of data is generated during this process and then comes the challenge like making insights out of this data where machine learning models come into play.
Many other sophisticated tasks are being augmented with machine learning. For instance, testing the weight of a potato before and after peeling. For this the R&D team is deploying deep learning models with classifiers such as Random Forests. These neural networks are fed with the images of the potatoes and are trained to give an approximation of the weight by identifying the percentage of peel. This eliminates any chance of extra peeling of the potatoes.
AI has found its way into almost all departments at Pepsico; from maintaining tastes to avoiding wastage. It has given control over what goes into the process and how good the product is. Not only that Pepsico has one foot in to the managing the output as well, with its Snackbot. Snackbot is a robot that delivers chips on a university campus. This driverless bot manages to steer through different obstacles by employing machine vision techniques.
It is being churned out for optimisation, no matter how small and, in domains never before imagined. The transition to data driven strategies is rapid. From product development to sales and marketing, there has been a tremendous improvement and companies are filling their pockets or at least saving them from spilling.
Check out other ways in which AI is put to use here.
Learn more about the photo-acoustic apparatus here.
- From Humanoid Robots to Telecom Trials, 12 Major Announcements at NVIDIA GTC Berlin 2026
- Anyone Can Make a Robot Move, Few Can Make It Stop
- After ISRO’s PSLV Failure, TakeMe2Space Looks to SpaceX for Reliable Launch
- Why AI Agents are the Next Frontier in Quantum Security
- Physical AI Squeezes Labour Market, And Low-Skilled Immigrant Jobs May Be First to Fall
- Rocket Math Isn’t Mathing: Why India’s Space Launches are Expensive