AI for the dairy industry at the EXPO

Suhas Patel, CEO of Tvarit, was at the EXPO to detail the “AI Co-Pilot for Dairy Industry For Sustainable And Zero Waste Production, Leveraging Deep Tech Software.” In his presentation, he asked, “Why can’t traditional optimisation methods solve today’s problems? Humans on the whole can track around seven variables using the five senses, but plants run with hundreds of parameters and so optimisation needs to understand all the parameters with many cheese products. These include coagulation, rennet and temperature, along with moisture retention, just to name a few. Cheese making is a very complex process. This is what we do with AI for optimisation,” he notes.

With the knowledge graph, the programme takes the data from the plant automatically, simulating the human way of working, and different areas are graphed using AI created by the software automatically. This creates the causal information and finds out why things are happening. The software helps the customers: the golden parameter values quality, anomaly detection and root cause analysis are all about optimising the variation range. It provides different optimum values and each cluster is a batch, with thousands of parameters monitored. Overall, using easy to understand dynamic generative AI powered prescriptions can improve the outcomes of a batch in a plant. Over the period of one year, one customer saw a 37% efficiency rate increase to 56%, by reducing moisture variation and increasing yields.

There has been adoption of the Tvarit process in the US market, and it is a growing trend in Austria and the Netherlands. The UK’s a new market, and the return on investment for the programme is fairly short, Patel observed.