Thank you to everyone who joined our recent webinar, From Prompt to Power: Automating Work with Quickbase AI Features. If you couldn't attend live or want to revisit the discussion, check out the FAQs below.
During the session, our product experts explored how customers can use the Quickbase Intelligence Pack to accelerate app building, automate repetitive work, and unlock insights while maintaining the security and governance you expect from Quickbase. We also shared a preview of upcoming capabilities. Read the recap and answers to questions asked live below.
Quickbase Intelligence Package
What is the Quickbase Intelligence Pack?
The Quickbase Intelligence Pack adds AI-powered capabilities to Quickbase that help builders, end users, and administrators work more efficiently. Current capabilities include:
- AI Agent
- AI Actions
- Data Analyzer
- App Intelligence
- AI Control Center
These features are designed to accelerate application development, automate repetitive work, surface operational insights, and provide enterprise-grade security and governance.
Who can use the Intelligence Pack?
The Intelligence Pack is available to customers who purchase the Intelligence Pack add-on.
Current Quickbase customers can activate a free 60-day trial from the Realm Administration page.
What AI capabilities are planned next?
The product team previewed several upcoming enhancements:
- Knowledge Layer – enables the AI Agent to answer questions using company documentation such as SOPs, manuals, SharePoint content, compliance documents, and onboarding materials.
- Schema Management – allows the AI Agent to update reports, dashboards, forms, relationships, and other application components through conversation.
- Workflow Agent – generates, validates, simulates, and prepares complete Pipelines for review before activation.
The first version of the Knowledge Layer is expected before the end of the year.
AI Agent
Can the AI Agent build Quickbase applications?
Yes. Alex demonstrated an application that was built entirely using the AI Agent, including:
- Tables
- Relationships
- Fields
- Sample data
The AI Agent can significantly accelerate initial app creation.
Can the AI Agent analyze reports and explain unusual values or outliers?
Yes—with some limitations. The AI Agent cannot currently analyze an existing report directly. However, it can:
- Query tables and apps the user has permission to access
- Join data from multiple tables
- Generate reports through conversation
- Explain unusual values or outliers in the returned data
Can AI Agent manage reports or dashboards?
The upcoming schema management capabilities will allow AI Agent to create and manage reports and dashboards through conversations.
Note: these can be created in the conversation, but can't yet be saved as Quickbase reports.
AI Actions
What are AI Actions?
AI Actions are Pipeline steps designed to automate repetitive manual work. They use AI to perform tasks such as:
- Extracting information from documents
- Classifying incoming requests
- Generating summaries
- Drafting resolutions
- Assigning work
- Creating structured outputs
Can AI input data into Quickbase?
Yes. AI-generated structured output can populate Quickbase fields through AI Actions. AI can generate values (such as assignments, summaries, classifications, or recommendations), and those outputs can be written directly into records as part of a Pipeline.
How do AI Actions avoid hallucinations?
Rather than relying solely on the language model, AI Actions can incorporate existing Quickbase data into prompts.
For example, Alex's demo searched previous closed requests and used those historical resolutions to draft a suggested resolution for a new ticket. This grounds the AI's response in organizational knowledge instead of generating an answer from scratch.
How are search results used during the Draft Resolution example?
The Pipeline:
- Searches previously closed requests.
- Filters for similar requests.
- Passes those historical records into the AI prompt.
- Uses those prior resolutions to draft a suggested resolution.
- Writes the draft back into the new request record.
This allows the AI to leverage previous organizational knowledge rather than inventing a solution.
Can you share more about the Query step used in the demo?
The query searched the Requests table and filtered records based on:
- Status = Closed
- Matching category
- Matching related products
Only those filtered historical records were passed into the AI Action.
How does AI classify request priority if users always write "URGENT" or "ASAP"?
We recommend including business rules directly within the AI prompt. For example, prompts can instruct the AI to:
- Evaluate the severity of the request
- Avoid relying solely on words like "urgent" or "ASAP"
- Score urgency based on the overall context
This allows organizations to prevent users from artificially inflating request priority.
Are AI Actions limited by document size?
The current supported file size limit is 3 MB. There are plans to increase this limit to 10 MB.
Is example YAML available for AI Pipelines?
Yes. The Quickbase Help documentation includes an example YAML definition for AI Actions pipelines.
Security & Governance
Does the Intelligence Package respect user permissions?
Yes, it honors existing Quickbase security, roles, and permissions. Users can only access data they already have permission to view, and AI responses are grounded only in that accessible data.
Are there safeguards to prevent accidental deletion of records?
Yes. If a user asks the AI Agent to delete records:
- The AI first verifies that the user has permission.
- The AI displays what will be deleted.
- The user must explicitly confirm before deletion occurs.
Can AI be enabled only for certain users or features?
Yes. The AI Control Center allows Realm Administrators to:
- Enable or disable individual AI features
- Grant access by user or group
- Roll out AI capabilities gradually
- Maintain governance and audit controls
Performance
Will AI Agents slow down Quickbase?
No significant performance impact should be seen. AI-generated queries are different from traditional in-memory formula queries and are not expected to introduce meaningful system slowdowns.
Integrations
Will Quickbase connect directly with Claude or ChatGPT?
Currently, A ChatGPT Pipelines channel is available. A Claude integration is under consideration, but a decision will not be finalized for a few months.
Licensing
Can I use the Builder Program to build AI proofs of concept?
No, the Builder Program does not currently include Intelligence Pack capabilities.
Forward Deployed Engineers (FDE)
What is a Forward Deployed Engineer?
Forward Deployed Engineers (FDEs) work directly with customers to identify AI opportunities and implement production-ready AI workflows.
Their work typically includes:
- Evaluating AI readiness
- Reviewing application schema
- Designing AI use cases
- Building AI pipelines
- Coaching customers on AI prompt design and implementation
Customers interested in working with an FDE should contact their Account Executive or Customer Success team.
Ready to dive deeper into the AI Discussion?
Join our App Builder Qrew Meetup on Tuesday, August 18 for a virtual discussion on how others are using the Quickbase Intelligence Package. You’ll hear real-world examples and use cases from customers marking the most of AI Agent and AI Actions.
Or, try the features for yourself. Start a free Intelligence Pack Trial >
