Skip to content

Why Continual


Machine learning and AI have the potential to transform every industry and every business function. But deploying AI today requires time-consuming manual effort for each use case, a team of highly skilled engineers and data scientists, and costly and complex machine learning infrastructure.

The result is an AI Gap — 95% of business leaders say their organization would benefit from embedding AI in their business, but only 6% report adoption of AI across their organization.1

This should be of no surprise. When every use case requires custom engineering, already overstretched data teams have no hope of keeping up. Even something as basic as maintaining a customer churn model can take months to make it to production. As a result, business leaders are frustrated by the lack of ROI from AI investments and data teams are overwhelmed by managing the accumulating technical debt of pipeline jungles.

The path to AI has led us from the era of Big Data to the era of Big Complexity.

No-code AI tools don't solve the problem. While promising increased accessibility and automation, these tools lack the operational characteristics that modern data teams expect, such as version control, the separation between development and production environments, and workflow automation.

There must be a better way.

What if you could build continually improving predictive models directly on top of your existing cloud data warehouse? What if you could make operational AI accessible to your entire data team by extending their existing analytic skills and workflows? What if you could leverage a declarative approach to AI — like SQL and dbt do for analytics — to eliminate complexity and ensure reliability?

That is what Continual seeks to deliver.

What makes Continual different?

Built natively for cloud data warehouse

Unlike traditional ML engineering platforms, Continual is built natively for cloud data warehouses like Snowflake, Redshift, BigQuery, and Databricks. Data-driven companies everywhere are rapidly moving to these data platforms and embracing SQL as the lingua franca of data. The result is a unified foundation for a new modern data stack.

Continual sits directly on top of your cloud data warehouses and automates the maintenance of both the predictive models to ensure they never go stale and the model predictions so your business always runs on up-to-date insights.

By writing directly back to your data warehouse, you can easily consume up-to-date predictions from existing BI, Reverse ETL, and downstream tools. There is no new infrastructure or complex integrations to manage to make predictions actionable.

Snowflake Quote

A declarative workflow for modern data teams

Unlike point-and-click AI platforms, Continual is powered by a declarative workflow built for modern data teams deploying production models across their business. At a high level, Continual operates in three steps:

While Continual does support UI-based development, it also has first-class support for dbt. dbt users can simply annotate existing data models to organize features into a shared feature store and automatically build and maintain new predictive models.

The result is a radically simplified approach to operational AI that allows you to focus on driving business impact from AI rather than infrastructure or operations. To understand how it all works, you can read more about Continual's core concepts and follow our quickstart.

dbt Quote

End-to-end automation for operational AI

Finally, unlike notebook platforms or simple ML-in-SQL extensions, Continual is operationally focused. It is designed for a world where you have dozens or even hundreds of models in production powering every aspect of your business from sales, marketing, operations, support, finance, and product. Building upon Continual declarative workflow, Continual maintains models and predictions with end-to-end automation, visibility, and governance baked in.

This focus on operational robustness, modern workflows, and complete visibility extends across the entire AI lifecycle from development to production. Using Continual, you can separate between development and production environments, leverage version control, and even integrate with CI/CD workflows all while avoiding pipeline jungles or complex engineering.

Operational AI UI

Next steps

To understand how Continual works in action, we recommend following our quickstart.

  1. Juniper Networks. 2021. AI is Accelerating, Is Your Organization Ready? 

Back to top