What is Cortex?

Cortex enables companies to integrate Strategic AI into their organizations quickly and easily, and optimize for their key strategic goals. Using Cortex businesses can typically begin showing the value of AI and machine learning in weeks vs. years.

Why use Cortex?

Cortex takes a lot of the pain out of deploying machine learning, automating everything from data ingestion, data engineering, module generation, and infrastructure for real-time APIs

screen-shot-2017-07-31-at-11-05-58-pmscreen-shot-2017-07-31-at-11-08-58-pm

Using Cortex in the UI : Module Order Optimization

This blog post illustrates how businesses can use Cortex to optimize the order of modules in the UI – uniquely for each user. The module order can be optimized using Strategic AI for business goals like “reduce churn”, “total time played”, “total transactions”.

Picking the right rows and the right order of rows in the UI can result in a better customer experience and increases in key metrics your business cares about like total play time, total transactions, etc. Cortex allows you to set the goals you care about using Strategic AI, while Cortex Queries allow you to quickly and easily automate your business for those metrics.

Example Cortex Query

screen-shot-2017-08-01-at-11-25-50-am

Cortex Connectors

Real-Time Adaptive APIs. The reason for using real-time APIs is that the UI should change in real-time based on the current behaviors of the user. Vidora’s real-time machine learning technology adapts immediately to changes in user behavior to ensure the UI is optimized for a user’s current preferences. Vidora’s APIs are fast and distributed globally, typically returning results in < 50ms.

Another example of a UI whose rows can be optimized uniquely for each user using Cortex.

Results

How much can you expect to increase engagement by implementing Cortex Queries for module optimization? Below are results in which our partner looked to optimize for CTR on a content landing page.

A possible ordering of modules in the UI. In this case Cortex was asked to personalize the subset and ordering of modules for each user. It’s a difficult problem to determine what modules to show each user given the hundreds of possible module categories available to choose from.

screen-shot-2017-08-01-at-11-18-49-am

Using Cortex increased the CTR by 40% and resulted in a more customized 1:1 experience for each user.

The performance increases, in terms of CTR, vary based on the number of rows available to choose from and the number of rows available to be shown in the UI. In general the more possible rows Cortex is selecting from and the more rows available in the UI, the higher performance gain one will see using Cortex.

Other Cortex Queries Relevant for Module Optimization

Cortex Queries provide a flexible framework for asking Cortex a variety of additional questions For instance you could adapt the query above by only optimizing module order for high activity users :

screen-shot-2017-08-01-at-11-21-35-am

Alternately you could restrict the modules shown based on the interest of the user:

screen-shot-2017-08-01-at-11-23-59-am

Finally you get set a different Strategic AI goal, like reducing churn:

screen-shot-2017-08-01-at-11-25-06-am

* Typically, for media companies, optimizing for churn reduction vs. CTR will result in content rows being shown which are biased towards episodic content. The idea being users get “hooked” on episodic content and come back week after week.

Your business is unique, Cortex Queries allow you to implement unique business logic to automate your business.

Want to Learn More About Cortex Queries?

Interested in leveraging Cortex Queries to optimize modules for your business? Email us at partners@vidora.com for a demo of Cortex Queries, and Cortex Insights.

Want to Learn More?


Schedule a demo and talk to a product specialist about how Vidora’s machine learning pipelines can speed up your ML deployment and ultimately save you money.