PrediQ implementation for a large chocolate manufacturing firm

Business Problem

  • Inability to correlate quality failures with out of sped process data
  • Reactive approach to handle quality issues resulting in time & inventory holdup
  • Quality out of spec is higher than industry standards


  • Using the existing data measure ,quantify and control drivers of Particle Size (PS – one of the Quality parameters)
  • Build a customized Machine Learning model to measure & proactively control process attributes within desirable range
  • Recommend equipment readings to bring particle size within spec based on simulation

Benefits to End Customer


Estimated 30% particle size defect

Picture1zxc (4)

Reduction in Cost of Quality

Picture1zxc (5)

Automated process with live