Data Productivity Cloud AI FAQ
This document provides answers to some frequently asked questions regarding Matillion's AI features in the Data Productivity Cloud.
Frequently asked questions
What do Matillion's AI features do?
Matillion incorporates pipeline components that allow customers to enrich their data, using third-party LLMs ("AI in the pipeline"). Additionally, Matillion provides productivity features with AI, such as auto-documentation for pipeline notation, and Copilot, which helps to build pipelines using natural language ("AI in the product").
What AI capabilities are available in Matillion Data Productivity Cloud?
- AI in the product: Improves the user experience by introducing automation, enhancing pipeline design, troubleshooting, and monitoring. This improves productivity and makes pipeline operations more intelligent.
- AI in the data pipelines: Lets users augment and enrich their data using third-party AI services integrated directly into Matillion pipelines.
Does Matillion provide its own AI processing capabilities for the AI Prompt components?
No. Matillion integrates with third-party AI services (AWS Bedrock, OpenAI’s ChatGPT, and OpenAI for Azure) to provide AI capabilities. By using these features, customers consent to the use of these third-party subprocessors.
Does Matillion use my data to train third-party AI models?
No. Your data is never used to train third-party AI models or to enhance Matillion's AI services. Matillion’s AI auto-documentation uses prompt engineering to translate our Data Pipeline Language (DPL) into business-readable formats.
In some cases, Matillion AI uses metadata (such as DPL prompts and pipelines) along with internal data (for example, documentation and engineering best practices) to generate pipeline suggestions. However, no customer data is accessed, and reinforcement learning processes do not utilize customer data.
Can customers bring their own AI accounts?
Yes. Customers can integrate with third-party LLMs via the AI Prompt components using their own API keys from providers including AWS, Azure, and OpenAI. Matillion doesn’t offer a built-in AI provider for these components.
Can customers opt out of Matillion's AI features?
- AI in the Pipelines: Users must manually add AI components to their workflows. If they choose not to use these components, the AI features remain inactive.
- AI in the Product: These features are available as part of Matillion’s Enterprise edition. Customers who don't use the Enterprise edition won't have access to these features, and Enterprise customers can choose to avoid using them. Visit Matillion pricing to learn more about each edition.
In both cases, there is no formal consent process. The choice to use AI features lies entirely with the customer based on their actions.
How is my data protected when using AI features?
Matillion ensures that no AI-generated data is stored within its systems. Secrets—including passwords—are not transmitted to AI services and are only resolved at pipeline runtime using secure methods. All data in transit is encrypted in accordance with Matillion’s security whitepaper.
What is Matillion’s commitment to responsible AI use?
Matillion’s approach to AI emphasizes augmenting human productivity rather than replacing it. Our goal is to empower teams to achieve greater success with AI-driven tools. Read our blog post here to learn more.
Where can I find more information about Matillion's security architecture and processes?
Visit trust.matillion.com for more detailed information about security measures and architectural processes.