Bakeries Add a Dash of AI and Sensors to Optimize Production

Bakeries heat up profits with a boost from artificial intelligence. Coupling a smart edge controller with cloud-based reporting results in perfectly baked goods every time, no matter the quantity or who is on the clock.

 

Article Key

Image credit: PreciTaste (Baking on Demand)

Families are expected to gather in smaller groups to celebrate Thanksgiving this year, but consumers likely will cling to traditional foods during the holidays to maintain a sense of normalcy. One shift will be that consumers are likely to purchase more pre-made items, including baked goods.

That’s great news for bakeries, if they can keep up with demand without over producing. Baked goods have a short shelf life, so having the ability to track purchases and adjust inventory on the fly is critical to avoiding food waste and burning profits.

The Smart Easy Bake Oven

PreciBake, the baking division of PreciTaste, is tackling this problem using artificial intelligence and predictive analytics. The company’s solution combines AI-enabled edge computing with cloud-based reporting to help commercial kitchens operate more efficiently and profitably. Powered by the Intel® Atom processor and the Intel® Distribution of OpenVINO™ toolkit, PreciBake’s Virtual Baker edge controller is mounted on a commercial oven.

The device uses an integrated Intel® RealSense™ depth camera to detect and identify bakery items as they are loaded into the oven. The Virtual Baker then automatically selects the appropriate bake cycle—including temperature and length of time—for that product.

Baking can be finicky. Ovens can be calibrated differently or form hot spots over its life cycle. Dough is susceptible to environmental factors, such as elevation, air temperature, and humidity, all of which can affect baking time. The Virtual Baker monitors hundreds of data points and uses machine learning to adjust as needed. It can identify and count products, offer suggestions to staff, and analyze operational data to instruct staff.

Because it is fully automated, the Virtual Baker takes the guesswork out of baking and ensures product consistency across employees and shifts. The autonomy translates to less time spent training kitchen staff, which often has a high turnover rate, and less food waste from user error. PreciTaste claims its clients have seen sales increase by an average of 24 percent.

The Virtual Baker stores baking data at the edge and sends it to PreciTaste’s TasteOS cloud reporting platform. TasteOS can receive data from multiple devices and locations. From there, management can gain insight about specific items, individual stores, oven functions, and overall performance across multiple locations. They can even see images from inside the ovens.

Commercial Kitchen Confidential

Armed with AI capabilities, commercial kitchens can go a step further and predict what items need to be baked and when. PreciBake’s Baking on Demand solution uses vision sensors to monitor inventory in real time and predicts demand every 15 minutes. It notifies staff what to bake, how much, and when to do so. The constant monitoring matches supply with demand to optimize inventory.

Croissant rolls baking temps are measured with sensors

Image credit: PreciBake

The Baking On Demand solution feeds into the BakeIT Cloud, a vendor-independent cloud-based reporting platform. BakeIT collects data from all the AI-enabled ovens, kitchens, and locations, allowing managers to view inventory, production, baking, costs, and other related data from a dashboard on a computer, smart phone, or tablet and adjust operations in any location.

PreciBake even addresses quality control with its Virtual Quality Officer. It monitors dough on commercial baking lines to ensure consistent size, form, and composition of the dough prior to baking. Machine learning capabilities can detect human intervention and enable the Virtual Quality Officer to identify flawed items before they are baked and become “seconds” or scrap.

As people gear up to buy rolls and pumpkin pie for Thanksgiving, bakers can use IoT to determine just how much to have on hand.