Skip to content
For Self-Managed AI Platform users running v11.0, see the on-premise platform documentation
Get started > First time here? > Fundamentals of generative AI (GenAI)

Fundamentals of generative AI (GenAI)

Generative AI (GenAI) available in Workbench, allows you to build, govern, and operate enterprise-grade generative AI solutions with confidence. The solution provides the freedom to rapidly innovate and adapt with the best-of-breed components of your choice (LLMs, vector databases, embedding models), across cloud environments. DataRobot's GenAI:

  • Safeguards proprietary data by extending your LLMs and monitoring cost of your generative AI experiments in real-time.
  • Shepherds you through creating, deploying, and maintaining safe, high-quality, generative AI applications and solutions in production, using your tool-of-choice.
  • Lets you confidently manage and govern LLMs in production, quickly detecting and correcting unexpected and unwanted behaviors.
  • Continuously improve GenAI applications with predictive modeling and user feedback.

This page provides an overview of the functionality.

For trial users

If you are a DataRobot trial user, see the FAQ for information on trial-specific capabilities. To start a DataRobot trial of predictive and generative AI, click Start a free trial at the top of this page.

GenAI overview

The DataRobot GenAI platform provides both API and GUI options, allowing you to experiment, compare, and assess the best GenAI components through qualitative and quantitative comparisons at an individual prompt and response level. The DataRobot AI Platform supports experimentation with common LLMs or you can bring your favorite libraries, bring or choose your LLMs, vector databases, and embeddings, and integrate third-party tools.

GenAI problem statement

With open source and GPU clusters offered by the cloud providers, it's possible to build GenAI workflows yourself, but it's hard to get those solutions into production in an enterprise setting. It's challenging to get the governance and guardrails in place to ensure you can confidently put your workflow into production while still ensuring that it meets your compliance and performance standards. With consultants that build an agentic application for you, you don't have the flexibility to adjust the implementation as your needs change over time.

While there is a perception that "it's all about the model," in reality, the value depends more on the GenAI end-to-end strategy. Quality of the vector database (if used), prompting strategy, and monitoring, maintenance, and governance are all critical components of success.

DataRobot solution

Sitting between do-it-yourself open source options and the full white-glove treatment of consultants that build GenAI workflows and agents for you, DataRobot provides:

  • An AI platform that allows you to build whatever you need to support your business with Generative AI.

  • A suite of applications that show you how to implement end-to-end solutions for common use cases.

  • One unified experience for all users, no matter the ingestion data layer, cloud provider, or consumption layer.

DataRobot is the "AI intelligence layer" that sits between the data infrastructure (unstructured data, on the left in the image below) and the consumption layer.

DataRobot provides a comprehensive suite of tools to build and evaluate GenAI workflows. Out-of-the-box observability and state-of-the-art inference let you scale workflows and understand what's happening in production and across the entire lifecycle. Governance functionality ensures you're comfortable with your workflow in the enterprise setting. Application templates provide pre-built applications for common use cases, as well as the code associated with the implementation. This allows you to customize—now and in the future—how all aspects of the application work, ensuring that the needs of your use case are met.

The general DataRobot architecture is used for both generative and predictive AI, whether it's preparing data, building and optimizing, or sharing insight. From experimentation in Workbench through deployment in Registry, all models can be managed and monitored through Console production capabilities. This includes governance, ability to deploy and manage, and continuous and ongoing monitoring.

Tip

See the generalized discussion of the steps to build generative models in a playground. Or, try it yourself with the GenAI walkthrough.

Read more