CoreNet Global Summit, Boston 2018
Boston | Oct 14 - 17, 2018 | John B. Hynes Veterans Memorial Conv. Ctr
Don't miss the largest gathering of corporate real estate professionals.

Lessons Learned from a CRE Machine Learning Proof of Concept

October 16, 2018 | 11:30 - 12:15

Lessons Learned from a CRE Machine Learning Proof of Concept

Innovating in the corporate real estate world brings with it many challenges that require patience, empathy, relationship building and an endless persistence in driving to a meaningful result. This session will look at a firsthand case study of a Machine Learning pilot that encountered all of the above-mentioned challenges and more. The project focused on leveraging machine learning to simplify and ease the work order request burden employees face, while automating some of the validation, ticket creating, and routing hurdles inherent in some of the legacy work order systems. The project leveraged Natural Language Processing (NLP), using just a phrase, word or sentence entered by the employee to create a real work order. The project was an innovation journey, facing both technical and cultural hurdles in creating a solution that was unknown, in a manner that was outside the norm of the typical CRE technology approach. Understanding what machine learning models are available, the process of data set gathering, and the data science viewpoints on interpreting the results were all aspects that will be valuable to anyone interested in leveraging a project like this. 

Key Takeaways:

  • Understand the fundamental need to identify where your critical data resides today, to consolidate and organize the data so it’s available for further analysis, and review the quality of that data. 
  • Understand the benefits of engaging enough stakeholders from all parts of the organization to avoid hurdles. 
  • Discover the value of learning from what you’ve done. Doing something completely new will cause some missteps along the way, but you’ll gain insights during the journey and become much more knowledgeable when finished.