Pilot electronic noses for quality testing

Food manufacturers detect taste and quality deviations with electronic noses

Future prospectAI Act: no extra obligationsMonitoring sensor and machine dataFood & consumer goods

Fits: Production, engineering & maintenance · Research & development

Odour and taste deviations, e.g. from off-flavours, spoilage or the wrong raw material, could be detected in minutes rather than by the sensory panel, including between tastings. So far, this mainly happens in research and pilot projects for individual products, not as a mature standard solution.

Is this for you?

  • You want to test, as a pilot partner or early in your company, whether electronic noses are worthwhile for a particular product.
  • A clearly described odour defect regularly causes you trouble, e.g. off-flavours in coffee, oil, fish or beverages, and the sensory panel cannot test every batch.
  • You have samples with known results from sensory testing or the lab that can be used to train a device.
First step today

Use the template below to research how far electronic noses have come for your product. Then work out how many samples with clear results from the sensory panel or the lab you had in recent months, good and defective. This shows you whether a pilot would have enough training material at all.

How to go about it

  1. Define the product and defect profile
  2. Set up sampling and measurement
  3. Train with good and defective samples
  4. Set thresholds for warnings and re-testing
  5. Test warnings in goods receipt and quality assurance
  6. Evaluate the pilot and decide

Templates to copy

Try it now
in the training
Analyse
in the training
Plan the rollout
in the training
Included in the training

The complete guide

  • 6 steps explaining exactly what to do
  • 3 proven templates to copy into your own AI
  • Safe use: data, checking, deciding, informing
  • Pitfalls, effort and how to recognise success

With the training you get all 430 guides for as long as your access runs.

See the trainingOne-off per person, no subscription

Same approach, other use cases

8 more use cases follow the pattern “Monitoring sensor and machine data”.

Detect battery degradation in your EV fleet earlyDetect damage to tracks and points earlyDetect falls in residents’ rooms without a cameraHave AI monitor process data and suggest setpointsMonitor cows with sensors and barn camerasPredict machine failures from vibration and temperatureDetect leaks in the water network earlyPilot network fault prediction and automatic control
‹ PreviousSee all use casesNext ›

LK-07 · Legal status 08.10.2026