Product

Industries

Resources

Company

Surface Inspection

Cake Inspection

Cake Inspection

Maddox AI reliably detects surface defects on cakes—such as burnt batter, smudged dollops, missing decorations, or dirty edges—at full line speed, ensuring every cake is inspected against consistent criteria.

Product

Product

Cakes

Industry

Industry

Food & Beverage

Total savings / year​

Total savings / year​

€100,000–€150,000 per year

Close-up view of a white cream cake with uniformly arranged cream rosettes along the edge, set against a bright, neutral background.

Customer Request

Customer Request

Our customer produces nearly 100,000 cakes per day in a three-shift operation. Previously, two operators per shift performed manual visual inspections from above to detect defects such as burnt dough, missing or smeared toppings, contaminated edges, and missing decorations. Due to operator fatigue and subjective evaluations, defect-free cakes were frequently rejected, while defective ones slipped through. This led to unnecessary scrap and increased the risk of customer complaints. The objective was to fully automate the visual inspection process while matching the accuracy of a consistently attentive inspector. To achieve this, Maddox AI deployed a camera system at the existing rejection station. Inspection results are transmitted directly to the line PLC, which automatically ejects defective cakes. The system was initially launched with over 10 cake varieties, with plans to expand to more than 100 varieties.

Before

Before

Previous Inspection System

Previous Inspection System

Previous Inspection System

Visual inspection was performed manually across 3 shifts by a total of 6 employees.

Subjective assessments and fatigue led to unnecessary scrap and increased the risk of customer complaints.

No digital recording of defect data for process optimization.

After

After

With Maddox AI

With Maddox AI

With Maddox AI

Visual inspection was performed manually across 3 shifts by a total of 6 employees.

100% of all cakes are inspected automatically, and defective units are rejected immediately.

Each cake is assessed against the same defect definition, encompassing over 30 defect categories.

Error data is recorded digitally and analyzed in the dashboard to identify root causes early on.

Subjective assessments and fatigue led to unnecessary scrap and increased the risk of customer complaints.

Each cake is assessed against the same defect definition, encompassing over 30 defect categories.

No digital recording of defect data for process optimization.

Error data is recorded digitally and analyzed in the dashboard to identify root causes early on.

100 %

100 %

100 %

Automatically Inspected Cakes

30+

30+

30+

Error Classes

~1 / Sek.

~1 / Sek.

~1 / Sek.

Testing speed

Want to learn more about our solution?

Try it free now!

Would you like to learn more about our solution?

Try it now for free!

Would you like to learn more about our solution?

Try it now for free!

Four colorful aluminum coffee capsules in copper, purple, green, and blue on a wooden surface.

With the use of Maddox AI, different capsule types can be reliably distinguished during a product changeover. This saves tens of thousands of capsules that were previously discarded as a precaution.

Food & Beverage

Stainless steel thermoforming packaging machine with a roll of clear plastic film and a touchscreen control panel.

Maddox AI ensures stable and reliable surface inspection in foils production, preventing downtime and guaranteeing delivery capability at all times.

Food & Beverage

White cosmetic bottles moving along an automated production line conveyor.

Maddox AI’s software, as well as the implementation of three individual camera systems, can reliably inspect surfaces, labels and imprints.

Consumer Goods