Loader

CASESTUDY

Avochato

Business texting is the new email

Using A.I. to Understand Customer Satisfaction

Avochato helps businesses more efficiently connect with their customers using messaging as the primary form of communication. Their goal is to help businesses evolve alongside their customers and to ensure that they are always getting the best service possible

 

Avochato partnered with Quantum Analytica to develop an automated strategy to understand customer satisfaction and sentiment. Utilizing their 200+ million messages across 40+ million two way conversations, we created and deployed an A.I. model to score customer satisfaction based on these messages so that a businesses’ best customer service representatives can be immediately directed to those users who require personalized attention

Our Solution

Custom Language Modelling

In order to understand customer sentiment, we built a custom language model that was able to understand the contextual information present in a conversation. This model was fine tuned specifically for Avochato’s data so that the model could understand emoji’s, slang, sarcasm, and two-way communication interactions.

Semantic Understanding

Measuring customer intent is more than labeling a conversation positive or negative. Using our custom language model, we created a novel machine learning pipeline so that Avochato could pinpoint where exactly in a conversation an interaction was trending positive or negative and the magnitude of that trend so that businesses could apply intervention techniques in real-time!

Scalable Deployment

Our model was deployed through an easily accessible API built in Django. The result was containerized in a docker image so that it could be easily scaled in a K8 environment

Automated Retraining

In order to provide maximum ROI, our model was designed to be automatically retrained on defined cadences so that it can always take into account

Let’s work on a project together