NLI Implementation In Automotive Retail And Aftermarket

This article is about the current Natural Language Interface implementations in Automotive Retail industry vertical, which mostly comprises of Auto Dealers.

In NADA 2018, Cox Automotive’s vAuto unit has showcased Nexi, which integrates with Amazon’s Alexa and is the first Artificial Intelligence powered digital assistant for auto retailers.

Nexi is capable of answering questions on dealer’s inventory, local market conditions, operations and more.

Nexi can respond to verbal requests for information or send alerts via text, email or voice.

Nexi implementation comprises of:

• an Natural Language Interface with voice that takes questions from staff of auto dealers on inventory and provides responses from the structured data – vAuto’s Stockwave in this case or in future could be from dealer’s DMS records
• a set of algorithms that trigger alerts based on occurrence of certain events that effect the data, like Stockwave offers to act on etc..

Whie Natural Language Processing technology is used in the Natural Language Interface with voice, triggering and sending alerts seems to be implemented without any Machine Learning.

Another company – Vast – has developed a Machine Learning based solution – CarStory – for Sales forecasting and triggering automated alerts.

CarStory endorsed by FordDirect is the first product to tell dealers when every car will sell, and for how much.

CarStory has captured and stored more than 200M listings. Using this information, a complete list of features and options were generated and a Machine Learning model was trained and built. This model is being used to predict or forecast Sales of dealer’s inventory. It is like providing Inventory Intelligence as opposed to Inventory Management!

Further the ML model is capable of learning itself from new data, like 10m listings that CarStory tracks everyday.

If we compare the above two implementations, Nexi has an Natural Language Interface implementation implementation using AI and an automated alerts without a Machine Learning model.

CarStory doesn’t (yet) seem to have an NLI for her Machine Learning model that predicts Sales of new/ used automobiles and also, sending automated alerts based on these predictions.

Now a days, we are seeing more and more customers using Alexa or Google Assistant or Cortana or Siri or other voice based NLI implementations to either access information or delegate certain tasks that can be automated to save time and money.

In this context, I see a definitive need for Auto Retailers to include in their Software Tools repertoire NLI implementations that include predictive analytics, like Sales forecasting engine, powered Machine Learning models.

Such NLI implementations should be scalable to include Aftermarket and be seamlessly integrated across assorted platforms, like organization’s Website or ERP or DMS or Social media etc..