DataRobot today added an enterprise edition of its platform for building generative artificial intelligence (AI) applications and associated agent software that adds a set of customizable templates to reduce the amount of time that might otherwise be required to less than 30 days.
While the use cases for generative AI vary widely, the core scaffolding for building and deploying generative AI applications is fairly similar, says Venky Veeraraghavan, chief product officer for DataRobot.
That commonality makes it possible to deliver a series of templates based on best practices that make it possible to deliver, for example, a data analysis application in 30 days versus waiting on IT teams to build and maintain a custom application built from scratch, he adds.
At the core of the DataRobot Enterprise AI platform is a GitHub repository that provides the foundation for sharing code across a platform-as-service (PaaS) environment for building and deploying AI applications. “It’s a GitOps-based approach,” says Veeraraghavan.
The templates being added extend that core platform in a way that even further reduces the amount of boilerplate code that an application development team would otherwise have to create while still making it possible to tailor security, business and implementation logic, says Veeraraghavan.
Initially, DataRobot is providing three templates that with the help of technology partners will soon be extended to six, he adds.
Other capabilities include rapid prototyping and deployment with automated monitoring tools included. Application development teams can also simplify maintenance by instantly pushing updates, fixes and improvements to applications without user downtime.
Application developers can build custom generative AI application interfaces with out-of-the-box examples for Streamlit, Flask and Slack, or create bespoke interfaces with their own preferred framework. A declarative framework for application programming interfaces (APIs) makes it simple to replicate work, visualize and save AI pipelines. DataRobot also provides an SAP Datasphere connector for integrating AI applications with the SAP AI Core platform.
Finally, IT teams can also stress test generative AI applications using a set of red team penetration tools provided.
Not every application development team prefers a PaaS environment that may limit the amount of flexibility they have to experiment with new tools, but the fact remains businesses are under pressure to deliver AI applications faster. The primary fear is that rivals will make use of AI to either reduce costs or increase revenues in a way that makes an organization substantially less competitive. In fact, IT teams can assume that if a template for a specific class of applications has been developed, it might simply be the new table stakes for remaining competitive. Not having that capability almost guarantees an organization will soon be falling behind. There may still be some justification for building an AI application development platform from scratch, but for many applications it just might not make any economic sense.
In the meantime, IT leaders can safely assume that business leaders will soon be losing patience with projects that appear to be taking a long time, require major investments, and have not shown any material benefit. As a result, it’s fair to say IT teams are under more pressure than ever to deliver on some aspect of the promise of generative AI sooner than later.
