Online Shop Logo SVG (White)

White Paper

White Paper

Originally published on February 14th 2023.
OSX Atom

The following white paper was composed and written by Terry McGinnis, Siraaj Ahmed and Dr. Torin Cannings PhD. The original concept of the overall engine has been in development since early 2017 and was published on June 8th, 2018 by Terry McGinnis.

The paper aims to present and explain how the Artificial Intelligence and Machine Learning engine, codenamed ‘OSX’, developed by Online Shop Inc., works and it’s applications for products and services readily available to the customer base of Online Shop Inc., it’s affiliates, stakeholders and partners.


OSX is being developed with commercial application in mind. Unlike other systems, learning models and engines which favor creative and linguistic approach, OSX is the first commercial and public Artificial Intelligence engine which aims to combine various learning models, data and language processing, computer vision, machine learning, big data and proprietary applications to automate and improve the quality of life of all users by replacing manual and tedious tasks by learning, adapting and overall optimizing every shop instance that is deployed via Online Shop, including any additional services and/or products.

The current engine is not readily available to the public as more data is necessary and scrupulous tests are required to ensure optimal operation in a live, non-test, environment.

OSX Big Data
What is OSX?

OSX is the codename for various Artificial Intelligence and Machine Learning models that encompass cognitive computing, creating a new definition of ‘value’ for those that use it. In the past such value was previously inaccessible or required significant resources to gather, analyze, understand and act upon. OSX is the next step of evolution, and an update to conventional computing based on mathematical principles.

Most systems use fundamental logic which is based on rules and logic intended to derive mathematically precise answers, often following a rigid decision tree approach. With advancements in technology and masses of data such rigid approach often breaks or fails to keep up with all of the available information, in a constant dynamic state that changes.

OSX enables users to create a profoundly new kind of ‘value’, finding answers and insights hidden away within clusters of big data with commercial intent. It provides a new approach to showcase context and derive action from volumes of data which is processed, to derive and provide value.

OSX Nucleus
The Engine

OSX serves to enhance human expertise by using it’s cognitive capabilities to mirror those of humans. It looks at problems the way a human might do. When humans seek answers, understanding or contemplate in the process of decision making they go through several steps to do so.

First they observe what is visible, and the environment which surrounds the observable. Secondly they draw on what they already know to interpret the visible, as a way to understand what the observable is, or what it means. Thirdly, humans evaluate whether what is visible is right or wrong which leads them to the final option, to act accordingly based on the best outcome deduced.

Just as humans become experts through the process of observation, evaluation and decision, OSX uses a similar process through its cognitive ability to reach reason from the data it collects. But unlike humans, it can do this at unfathomable speed and scale.

OSX Cube

Unlike conventional approaches of today which are only able to handle structured data in neatly compact databases, the OSX engine can parse and understand unstructured data. Such data is usually produced by humans for other humans, this includes everything from written content such as articles, posts, blogs and reports to differentiating expressions that are only understood by humans.

OSX relies on natural language and dynamic data models which stem from context, affinity, grammar and expression. Conventional computing relies on well structured data that is well defined and specified.

The OSX engine is able to parse through implicit, ambiguous, complex and challenging layers to process and understand such data to compose contextual actions that are backed by sets of data which is interpreted through the cognitive lens of a human. It achieves this by breaking down language, grammar and understanding the sentiment behind it, which it then combines with other data to structurally discern the overall meaning of input.

Unlike crawlers used by search engines of today which use keywords and synonyms to bring up results, OSX understands context. It tries to understand the real intent of a user to extract the best logical response to the intent. This is achieved through an array of language models, computer vision and algorithms.

When a user enables OSX upon deployment of their shop instance via Online Shop, it learns the language, the jargon and usage of the instance through the user. It then proceeds to collect behavioral data to achieve contextual analysis through the thought process and action taken of every visitor that visits the shop instance.

OSX aims to evaluate user behavior through heat mapping and event triggers, when combined with everything else, it provides a way for it to consider the best course of action to improve the overall visitor experience for that shop instance. It builds a lexicon of information from which it culls the most applicable and discards anything that is underperforming or irrelevant to that visitor.

As more and more data is collected and information is provided to the overall lexicon, it can use the aforementioned to improve itself via machine learning, allowing it to be grounded in truth based on data alone. It continues to learn through ongoing interactions and events, allowing it to constantly adapt and optimize, not only the deployed shop instance but itself. As more shop instances are deployed more data is submitted to the overall OSX training model, helping everyone achieve greater results.

OSX Engine

The ability to collect data and process information within contextual means allows for the OSX engine to identify new insights, visitor patterns which otherwise would of been locked away. It provides OSX to make new developments through evidence based decision making and analysis.

Unlike most models and systems, it makes decisions which are grounded in evidence to either support or refute it’s decision. It uses statistical modelling to assign a score, both for internal optimization and external for improvement of the overall visitor experience on a deployed shop instance, aiming to achieve a perfect Net Promoter Score.

The scoring system allows it to ensure a high confidence level in it’s decision making, allowing it to run it’s very own analytics to see insights to make the best and most informed decision. It uses the same analytics and event capture model which is available to users via Online Shop Analytics, a GUI and API that allows users to have a more advanced view of not only their shop instance performance but any other web property.

Most important of all, the OSX engine was designed to learn, adapt and get smarter with each new shop instance deployed and it’s interactions with users and visitors, gaining value with age.

Closing Notes

The OSX engine is still in development and much more training data is necessary to perfect it’s overall learning and decision making model and processes. We will not release an unfinished product, and aim to perfect it before making it readily available to the general public.

However, we are working with both governmental and private institutions to help improve it, and in turn help them with their own efforts.

Please note, that OSX only collects data from shop instances. Those who use the Online Shop Analytics offering on independent web properties do not have their data collected unless enabled by the user to send such data and information to help improve and train the overall model. If enabled, the data will be used to train the OSX engine on off-site visitor behavior to help it better understand and make more informed decisions for shop instances deployed via Online Shop and it’s supporting products, including affinity modelling, audience creation and other proprietary efforts. 

OSX Pathfinder
Additional Reading

Below are links to both internal and external reading material for those looking to gain more insight and a better understanding of the OSX engine, ideas and thought leadership which relates to the overall effort. The list may be updated periodically without notice.

  1. Terry McGinnis – Achieving technological singularity via Artificial Intelligence and Machine Learning – The vision for Online Shop
  2. Terry McGinnis – Expanding on ‘How it works?’ – an explanation of how Online Shop uses Artificial Intelligence and Machine Learning
  3. Terry McGinnis – IBM and the Next Generation of Advertising
  4. Terry McGinnis – Building the Next Generation of E-Commerce
Online Shop Protective Arms (Black)

We use cookies to offer you a better browsing experience.