Workrede provides an app-based solution for “non-desk” employee workplaces, such as restaurants or healthcare facilities. Their customers were interested in gaining insights from employee conversations to improve communication and identify problematic tone and sentiment. They needed to know first what was possible and approached Slingshot to explore the problem using Artificial Intelligence.
Workrede
A communication platform company that provides communication tools to the “non-desk” workplace.
The Story
Technology: TensorFlow, Machine Learning, Natural Language Processing
The Challenges |
The Solutions |
| Testing their idea at a lower cost relative to a full project was a key concern | Developed a rapid Proof of Concept, providing exploratory results quickly and reasonably |
| Massive amounts of conversation data needed to be analyzed and rated | Built a model that scans a conversation and is able to detect sentiment, emotion, and potential confusion over a process |
| Needed to make sure sentiment analysis factored in the specifics of an industry | Discovered that sentiment differs in different workplace environments, varying based on industry, adjusted POC accordingly |
The Challenges |
The Solutions |
| Testing their idea at a lower cost relative to a full project was a key concern | Developed a rapid Proof of Concept, providing exploratory results quickly and reasonably |
| Massive amounts of conversation data needed to be analyzed and rated | Built a model that scans a conversation and is able to detect sentiment, emotion, and potential confusion over a process |
| Needed to make sure sentiment analysis factored in the specifics of an industry | Discovered that sentiment differs in different workplace environments, varying based on industry, adjusted POC accordingly |
Hitting The Target
The proof of concept project wrapped up with Slingshot presenting the results of its sentiment analysis on the conversations to Workrede with a live demo. The analysis confirmed that sentiment could generally be detected, but it also highlighted challenges in interpreting sentiment across different industries as well as specific business process breakdowns.
This proof of concept allowed Workrede to test their ideas without building an entire app to do so. Workrede is excited about the possibility of integrating more advanced AI capabilities into its platform. This POC showed how AI could add value in analyzing workplace conversations and identifying areas for improvement.
