Sales Call Analysis/Predictions Using DS & NLP - BC-797

Project type: Research
Desired discipline(s): Computer science, Mathematical Sciences, Mathematics, Statistics / Actuarial sciences
Company: Replayz Labs LTD
Project Length: 6 months to 1 year
Preferred start date: As soon as possible.
Language requirement: English
Location(s): Vancouver, BC, Canada; Canada
No. of positions: 4
Desired education level: Master'sPhDPostdoctoral fellowRecent graduate
Open to applicants registered at an institution outside of Canada: No

About the company: 

Replayz helps top tier software organizations improve sales by using their call data to show them how to conduct more effective sales calls.  We do this by analyzing their recorded sales calls, scoring their calls, and using Data Science to isolate exactly why their win deals.

Today this is a tech enabled service, where our data science models ingest call scoring data, scored by human call scorers (data labellers) and we serve up valuable insights showing why their deals close and what their top reps do well on a sales call.

This project is about testing whether we can automate most of this process using statistical analysis and NLP.

Describe the project.: 

Innovation:  Replayz currently provides a specific set of capabilities as a human-intensive tech-enabled service. This project will allow Mitacs interns to develop a parallel pipeline of statistical analysis and natural language processing to recreate the output of the legacy process using a large set of labeled data. This new process will:

1) ingest a written call transcript for analysis by a large language model (i.e. GPT-J)
2) answer specific standardized questions about the contents of the transcript
3) perform statistical analysis on this dataset. The goal will be to understand how successfully a combination of large language models and traditional statistical analysis can perform on a prediction task vs existing methods rooted in human intuition. 

Success in this project will serve as a proof of concept for automated professional education systems using real-world examples of employee interactions.

Required expertise/skills: 

1) Knowledge base: Firm understanding of statistics and multivariant analysis.  Familiarity with generative AI methods and natural language processing.
2) Skills: Strong Python skills, experiencing integrating with external API’s and performing data analysis using standard python libraries.