--- title: "Quickbase AI models and security" slug: "quickbase-ai-models-and-security" updated: 2026-06-01T18:15:31Z published: 2026-06-01T18:15:31Z canonical: "help.quickbase.com/quickbase-ai-models-and-security" --- > ## Documentation Index > Fetch the complete documentation index at: https://help.quickbase.com/llms.txt > Use this file to discover all available pages before exploring further. # Quickbase AI models and security > *Access to this feature can change based on your Quickbase plan. Learn more about feature availability and plans in*[*Quickbase capabilities*](/v1/docs/quickbase-capabilities)*.* Quickbase AI is a collection of AI-powered features that empower you to work faster by simplifying complex tasks like creating apps, building pipelines, generating reports, and more. Quickbase AI features include: - Ask Quickbase AI - AI Smart Builder reports - AI Smart Builder apps - AI Smart Builder pipelines - AI Actions pipelines channel - Data Analyzer/Quick Insights - AI Data Scanner - Quickbase AI Agent We prioritize safety, security, and transparency in every Quickbase AI feature while maintaining Quickbase platform standards outlined in our existing [Shared Responsibility model](https://www.quickbase.com/security). ## Quickbase AI principles **Fairness and inclusivity** We use the latest models from industry leaders for Quickbase AI features and only select models regularly audited for potential biases to ensure equitable performance across all user groups. **Transparency** Quickbase AI features are clearly labeled in the platform and Quickbase AI interactions are logged in audit logs accessible to realm admins. **Continuous improvement** We continuously improve the models used for Quickbase AI features based on research, industry advancements, and your feedback to meet the highest standards for AI tools. **Privacy and security** Quickbase maintains industry-leading security practices across the platform and Quickbase AI features are governed under the same security standards to ensure privacy and security for your data. All AI model providers have gone through extensive reviews from our security, legal, and compliance teams. > [!NOTE] > Note: Quickbase never trains AI models on your data and your data is not retained by any Quickbase AI technology providers. When using Quickbase AI features, do not include any sensitive or personal data in your requests. ## Quickbase AI models Quickbase uses AI models to power Quickbase AI features. These models are selected and updated to ensure they are fair, inclusive, and secure. | **Quickbase AI feature** | **Description** | **AI technology** | **AI model** | **Quickbase data used** | | --- | --- | --- | --- | --- | | AI Smart Builder apps | Generate a Quickbase app by describing your industry and details about what you'd like your app to do. | Generative AI | Anthropic Claude 3 Haiku (hosted in AWS Bedrock) | User prompts | | AI Smart Builder reports | Create AI-suggested charts based on the structure of your table. | Generative AI | Anthropic Claude 3 Haiku (hosted in AWS Bedrock) | - Table schema data - Field metadata (Names, Types, Multiple Choice Values) | | AI Smart Builder pipelines | Create a pipeline | Generative AI | gemini-2.5-flash-lite | - User prompts - Channel connection data - Pipeline schema - Field schema - Select options | | Add steps | gemini-2.5-flash-lite | | Populate fields in a step | gemini-2.5-flash-lite | | Fill in field mappings in a pipeline | gemini-2.5-flash-lite | | Generate custom Jinja | gemini-2.5-flash-lite | | Summarize a pipeline | gemini-2.5-flash-lite | | AI troubleshooting in Pipelines | Generate explanations of pipeline errors and suggested solutions | Generative AI | gemini-2.5-flash | - Channel connection data - Pipeline schema - Field schema - Select options | | AI Actions pipelines channel | Generate an AI response within a pipeline based on the instructions you input | Generative AI | Amazon Nova Lite | User prompts | | Data Analyzer/Quick Insights | Use machine learning to analyze Quickbase record data and draw correlations within your data to create predictive insights. | Machine learning | ML including Logistic Regression within scikit-learn, statsmodels, and scipy* All AI technology for AI Data Analyzer is hosted within our cloud instances. No customer data will be sent to any 3rd parties for training purposes. | Quickbase record data | | AI Data Scanner | Identify sensitive data within apps using AI and machine learning techniques. | Machine learning | ML models used within MSFT Presidio All AI technology for AI Data Scanner is hosted within our cloud instances. No customer data will be sent to any 3rd parties for training purposes. | Quickbase record data | | AI App Intelligence | Get a summary of app metadata (structure, components, and key app information) | Generative AI | Claude 3 Haiku (hosted in AWS Bedrock) | Quickbase app schema | | Ask Quickbase AI for help | Ask questions about the Quickbase platform or the task you want to complete and get help resources. | Generative AI | Anthropic Claude Sonnet 4.6 (hosted in AWS Bedrock) | User prompts | | Quickbase AI Agent | Build apps, tables, relationships, and fields, ask for help with product and task questions, and analyze data. | Generative AI | Claude Haiku 4.5 Claude Sonnet 4 | User prompts Quickbase app data |