PRESS RELEASE
FOR IMMEDIATE RELEASE: 28 October 2026
Finnish AI company MarshallAI, backed by a growth investment from Bittium in 2025, is setting its sights on international beverage markets. The company’s technology can distinguish between nearly identical cans on high-speed production lines and inspect the content of date codes on every can at speeds exceeding 120,000 cans per hour.
TURKU, FINLAND. Finnish machine vision and sensor intelligence specialist MarshallAI is targeting international growth in the brewing and beverage industry. In November, the company will showcase its technology in Nuremberg, demonstrating quality inspections at high production speeds that have previously been extremely difficult to automate reliably.
MarshallAI’s Rogue Can Detector identifies incorrect cans that have accidentally entered a production line at speeds exceeding 120,000 cans per hour.
The problem sounds simple, but it is surprisingly difficult for machines to solve.
The same production line typically handles multiple beverages, and even a single product may have several nearly identical packaging variants for different export markets. The difference between two cans may be as subtle as a small piece of text, a different language, a symbol, or an individual packaging marking.
“Breweries have told us they’ve been struggling with these challenges for years. Production lines are getting faster, product ranges are expanding, and batch sizes are becoming smaller, resulting in more frequent changeovers. Our technology can inspect and distinguish even highly similar packaging variants in real time, without slowing down production.”
Marcus Nordström, CEO, MarshallAI
Conventional machine vision works well when the defect being inspected can be precisely defined in advance. Deep learning-based systems, however, can learn to recognize the entire package and detect certain types of defects they have not previously encountered, even at full production speed. The technology can also distinguish between nearly identical domestic and export versions of the same can.
A wrong can entering a filling line is rare, but it represents a serious quality risk. If a can from a previous production batch remains on the line during a changeover, the wrong beverage may end up in packaging intended for another product. This can lead to product recalls, rejected export shipments, and significant reputational damage.
AI does more than detect a date code – it checks whether the date makes sense
Another solution developed by MarshallAI addresses something familiar to almost every consumer: the date code printed on the bottom or side of a package.
Existing inspection solutions can check packaging markings, but often the inspection only determines whether a marking is present.
MarshallAI’s system also understands the information contained in the marking.
The system reads the best-before date, production date, and time from every package. In real time, it verifies whether the characters form a valid calendar date, whether the production timestamp matches the actual production time, and whether the best-before date falls within the expected shelf-life window defined for the product.
At the same time, the AI evaluates the readability of every individual character.
This enables the system to detect cases where a date code is present but contains an impossible or incorrect date, or where some of the characters are unreadable.
“We believe the most exciting potential of AI in industry is not simply automating human work, but enabling inspections that neither humans nor conventional machine vision can perform at production speed. With date codes, for example, we can simultaneously verify the readability of every character and check whether the dates and production timestamps are logically valid.”
Marcus Nordström, CEO, MarshallAI
Bittium invested in the Turku-based company in 2025
MarshallAI has been developing deep learning and sensor analysis solutions since 2014. Development of the company’s proprietary AFSA platform (Agile Focused Signal Analysis) began in 2017.
In 2025, the company’s growth ambitions gained the support of a major Finnish industrial partner when Oulu-based publicly listed technology company Bittium invested in MarshallAI. The two companies have since strengthened their collaboration in AI and software technologies.
In addition to industrial applications, MarshallAI’s technology is used in defence and security solutions. The beverage industry represents the company’s next major target for international expansion.
MarshallAI’s systems operate locally at production facilities using edge computing. This eliminates the need to transmit image and sensor data to the cloud and enables inspection decisions to be made in milliseconds directly at the production line.
Next stop: Nuremberg
MarshallAI will showcase its Rogue Can Detector and AI-powered date code inspection technology at BrauBeviale 2026 in Nuremberg, Germany, on 10–12 November 2026.
The company aims to establish new partnerships with breweries, beverage manufacturers, and technology providers to support its expansion across Europe and other international markets.
Visitors can meet MarshallAI at Hall 10, Stand 10-139.
Further information and interview requests
Marcus Nordström
CEO
MarshallAI / Kradient Intelligence Oy
Email: marcus@marshallai.com
Phone: +358 50 506 7329
Website: https://marshallai.com
Product demonstration video:
https://vimeo.com/1237528499
About MarshallAI
MarshallAI is a Finnish deep learning and sensor intelligence specialist based in Turku, Finland. The company develops AI solutions built on its proprietary AFSA platform, enabling real-time analysis of image and sensor data in demanding operating environments.
MarshallAI provides locally deployed, turnkey quality assurance systems for industrial applications, as well as AI and sensor intelligence solutions for customers in the defence and security sectors.
