Conference Day Two: Wednesday, 27th February 2019
Wednesday, February 27th, 2019
12:00 PM Blending Machine Learning And Image Recognition To Automate Student Verifications For Mobile Offers, Improve Satisfaction & Drive Back Office Efficiency
At Spark NZ, not only has the business embarked on a massive restructure towards agile principles but now has also seen the automation team venturing into innovative uses that move beyond basic RPA. Hear how the Automation Tribe conceptualized, iterated and delivered a robust application model that leverages ML and image recognition to reduce a verification time from 12 hours to 1-2 mins as well as save manual labor through an accuracy rate of 88%!
- Using Google’s TensorFlow framework to develop a ML program that operates at high accuracy
- Lessons on how best to plan for and deal with exceptions as you scale such application models across the business
- Overcoming challenges with the OCR service to improve image recognition and standardization