Why The Cognitive Engine?

6+ years were spent in academic institutions within the US, UK, Europe & China prior to successful commercial exploitation.

Chatterbox’s Cognitive Engine moves the goalposts, enabling partners to effortlessly leverage Artificial Intelligence, Machine Learning & Deep Learning capability without being Data Scientists.

Chatterbox’s proprietary Artificial Intelligence, Machine Learning & Deep Learning classifiers uniquely harness the power of statistical machine learning methods. These are the only proven methods for real-time data processing at scale.

The Cognitive Engine is made of constituent components that define processes and methods for agile development of new Artificial Intelligence, Machine Learning & Deep Learning products within days. Within the Engine we have training graphs, automated parameter selection, statistical ML, classification, real-time prediction engine & testing.

Build new products and use-cases in days.

Dynamic Learning

Drawing on Cognitive Science across Artificial Intelligence, Machine Learning and Deep Learning, Chatterbox Labs differentiates by using a process called Dynamic Learning to build statistical models of both text and numerical data. On the text side it is imperative that any technology recognizes both formal and slang language, industry specific nuances and mixed language texts without having to craft and maintain expensive and slow rules. The Cognitive Engine doesn’t use keywords, dictionaries or boolean queries - instead it achieves world leading performance by training the computer to automatically and continually learn from text and numerical data.


Source Code Deployment

The Cognitive Engine is coded in two languages: Python, which provides an extensible and customizable environment for experimenting, testing and training production ready Artificial Intelligence, Machine Learning & Deep Learning classifiers; and Java, which provides a real-time prediction engine optimized for massive parallelization.

This is not offered as a Software as a Service solution, instead partners have full access and modification rights to the source code, deployed and updated via a private repository. This enables partners to ensure privacy of client data, scale without latency and modify the codebase to align and integrate with other internal initiatives.

Real World Solutions

Cognitive Engine Real World Solutions

Speed, Accuracy & Volume

148k messages classified per second on 9 threads. 36k on a single.
600k Times faster than people at classifying data.
100% stateless so you can parallelize
83-91%  Accuracy
Setup within your firewall within 48 hours
Deployed in the cloud within minutes

The Cognitive Engine

Chatterbox's Cognitive Engine encapsulates the Dynamic Learning process. The engine has 6+ years of IP allowing for training graphs, automated parameter selection, statistical ML, classification, real-time prediction and testing without the need for Data Scientists or code level changes. Partners have the ability to leverage the Engine to build new Artificial Intelligence, Machine Learning & Deep Learning products for new industries, use cases and languages to address complex business scenarios.


Partners are empowered to realize the true value of their own data. Having access to source code and modification rights ensures data privacy, minimizes latency and retains in-house control.

Chatterbox is 2 years ahead of the curve and has an extensive roadmap to boot.