Trax
Active

Image Recognition for Retail Analytics and Points of Sale

Retail Visual Monitor
Artificial Intelligence Computer Vision Machine Learning

Business Overview

DESCRIPTION

Trax Image Recognition provides cutting-edge image-recognition technology and market-data services to tier-one manufacturers. The company assures tighter in-store execution controls and the ability to unlock revenue opportunities at all points of sale. Trax’s technology can be integrated as a plug-in to existing merchandising or retail execution systems. The company’s platform utilizes iOS, Android, and Windows smartphones and tablets to capture images of retail shelves. Analysis within the Trax cloud provides immediate, actionable mobile reports to sales reps and aggregated web reports to management teams. Trax’s solutions replace manual auditing methods by up to 60%, allowing sales reps to focus more time on in-store sales development activities. Trax is a trusted OEM partner of more than 25 leading retail solution providers, and many top brands use Trax to manage their in-store execution and increase revenues, including Coca-Cola, AB InBev, Nestle, and Henkel.

FOUNDED
November 2010
EMPLOYEES
335
BUSINESS MODEL
B2B
OFFERING TYPE
Software
BUSINESS STAGE
Launched
TOTAL FUNDING
$392.8 Million
SECTORS
Retail Visual Monitor

Funding Rounds

Date Announced

Funding Round

Amount Raised

Investors

July 2019
Series D
$100 M
July 2018
Private Equity
$125 M
June 2017
Private Equity
$64 M
February 2017
Series E
$19.5 M
June 2016
Series D
$40 M
December 2014
Series C
$15 M
February 2014
Series B
$15.7 M
August 2013
Series A
$6.6 M
June 2011
Seed
$7 M

AI Technology Stack

AI DESCRIPTION

We use AI for improving our own decision-making processes, for improving our products, and as a driver for new products. Internally, we leverage Time Series analysis (mostly Holt Winters Triple Exponential Smoothing or ARIMA) to predict the expected workload (images, analysis times and costs as well as identifying system performance issues (anomaly detection using GMM). AI is also ingrained in our products: we use non linear regression to predict our customer sales performance, Auto Encoders and NMF to create dense representation of outlets that are then fed into unsupervised clustering methods (often density based such as DBScan), and classification to build profiles of these outlets. Prediction results are often fed into planning and optimizing tools to generate concrete actions. These include heuristic search (A*, AKA œA-Star), Genetic Algorithms, and local search with simulated annealing.

AI EMPLOYEES
31
AI APPLICATION
Vertical AI
AI TYPES
Artificial Intelligence Computer Vision Machine Learning
ML TYPES
Supervised Learning Unsupervised Learning
AI ALGORITHMS
A* Agglomerative ARIMA Auto Encoders Boltzman Machines Boosting Clustering DBScan Decision Trees Density Estimation Genetic Algorithms Holt-Winters K-Means Local Search Naive Bayes Classifiers Nearest Neighbor Neural Networks Non-Linear Non-Negative Matrix Factorization Simulated Annealing Support Vector Machines t-SNE
AI TOOLS
Keras Scikit-Learn TensorFlow
CODING
Python
AI HARDWARE
CPU GPU

In The News