Airtable AI

What is Airtable AI?

Airtable AI represents the intelligent capabilities integrated into Airtable’s relational data and no-code application platform. The Omni interface serves as the main entry point, allowing users to create or modify Base structures, tables, fields, views, interfaces, and automations using natural language; it also enables users to query data, analyze attachments, research web pages, and update records.

Field Agents utilizes AI as a field that can be operated in batch mode or triggered automatically, to classify each record, generate summaries, extract information, create images, conduct web research, or build prototypes. It allows AI to work directly on structured business data, rather than producing chat responses that are disconnected from the database.

Core functions

• Omni application setup:Create tables, relationships, fields, views, and interfaces in natural language, modify existing applications or generate formulas; building such applications does not consume any AI credits.
• Data Q&A and operations:Querying records, summarizing trends, and creating or updating data – all of these actions are carried out within the scope of the permissions granted to the current user in Base and Interface.
• Field Agents：Following the prompts for each record, it can automatically classify feedback, extract documents, generate text, translate content, evaluate leads, and add structured fields.
• Document and image analysis:Read the text and information in attachments, extract summaries and fields, and it is also possible to use image generation or visual models.
• Web research:When networking is allowed, search for and retrieve public information, then record the results; the text obtained from such searches also contributes to the AI’s score.
• Automation:The AI is invoked when a new record is created or its status changes, after which an email is sent, a task is created, other tables are updated, or external applications are triggered.
• Prototype generation:Field Agent can generate interactive prototypes in React, Next.js, or HTML/CSS from descriptions or screenshots, and these prototypes can be previewed or deployed on Vercel.
• Models and governance:Administrators can enable specific AI platforms, view quotas, set workspace permissions, and enjoy enhanced management capabilities in both Business and Enterprise versions.

Price and AI points

AI points are shared collectively within a workspace or organization, and they do not represent an individual balance for each user. The Free, Team, and Self-Service Business plans have their points updated at the start of each billing period, while the Sales Management Business and Enterprise plans have their points updated at the beginning of every month.

Points expire each month and are not carried over.

Different tasks consume resources dynamically based on the input data, output length, model used, and network connectivity; advanced models, large attachments, in-depth research tasks, and image processing tasks require more resources. A warning is sent when the system’s usage reaches 75%, and the Owner or administrator can view detailed information in the settings.

Platform, API, and open-source status

Airtable offers web versions as well as applications for Windows, macOS, iOS, and Android; Omni and the full Field Agent setup are primarily used via the web and on desktop devices.

The platform also offers REST APIs, Webhooks, Extensions, and developer tools to integrate Base data with external systems.

Permissions, API rates, and record limits vary depending on the plan.

The Airtable platform, Omni, and AI model orchestration tools are not open-source software. The official GitHub repository provides SDKs, examples of Blocks/Extensions, and development tools, which are intended solely for extending the hosted services; they do not constitute a complete database platform or an AI backend.

The generated web prototype code can be further developed, but this does not mean that Airtable itself can be hosted on one’s own server.

Airtable AI usage guide

Complete a basic task.

1. Clarify the purpose of the image, its dimensions, style, and requirements regarding commercial licensing.
2. Upload the original images for which you have usage rights to Airtable AI, or enter clear instructions;
3. Choose the Omni application to set things up and generate low-cost candidates first;
4. Use data-driven Q&A and adjustments to refine the composition, subject, and details;
5. Zoom in to examine faces, hands, text, edges, logos, and artifacts;
6. Export after confirming the color, resolution, and licensing, while retaining the original image;

Create reusable professional workflows

1. Create a list of brand colors, fonts, layout elements, and elements that are prohibited.
2. Test the Omni application setup, data query and operation functions, as well as Field Agents separately;
3. Use the same set of representative samples to compare quality, speed, and cost;
4. Complex edges, text, and images of key products should be handed over for manual refinement.
5. Standardize naming, dimensions, and review status;
6. Batch processing and release are carried out after random inspections;

Which users are it suitable for

• Business teams that manage content, products, projects, customers, and operational data
• Researchers and marketers who need to batch-classify, summarize, and enrich records
• Non-technical users who wish to build relational applications using natural language
• Organizations that need to integrate AI into data permissions, interfaces, and automation processes

Restrictions and safety precautions

• Omni will perform operations on real data within the existing permissions; before carrying out batch updates, it should first outline the plan, limit the scope of the views, and create snapshots.
• Automatic execution of FieldAgent can quickly deplete points due to loops, large attachments, or a large number of records; it is therefore advisable to start with a small number of test records and set appropriate conditions.
• Web research and model outputs may be incorrect or outdated; it is necessary to verify them before entering data in the fields related to customers, prices, compliance, and finance.
• Attachments, hints, and records may be processed by the selected AI platform; administrators should only enable suppliers that have passed organizational approval.
• Sensitive data should be handled under the principle of least privilege, using corporate settings and appropriate data processing protocols;
• Web prototypes generated by AI are suitable for testing interfaces, but they are not equivalent to production applications that have undergone security audits;

Frequently Asked Questions

Can all of Airtable AI’s solutions be used?

Yes, all plans include a different number of monthly AI credits; once the credits in the Free plan are used up, an upgrade is necessary, and paid plans allow for the purchase of additional packages.

Will using Omni to set up a Base result in point deductions?

Currently, creating and modifying the application structure itself does not cost any points; activities such as Q&A, research, Field Agents, AI automation, and content generation do require points.

Does Airtable AI support Chinese?

It is possible to ask questions, categorize content, translate it, and generate text in Chinese, but the field rules, proper nouns, and the results of bulk modifications still require manual verification.