Airtable AI Governance After the Acquisition: What Enterprise Teams Should Review

Bending Spoons’ agreement to acquire Airtable comes at a time when AI is becoming a much bigger part of the Airtable platform.

Airtable now supports AI-powered app building, Field Agents, Omni, multiple AI model providers, and AI-driven workflows across business data. For enterprise teams, that creates opportunities but it also makes Airtable AI governance increasingly important.

The acquisition is expected to close by the end of 2026, subject to regulatory approvals and closing conditions. Until then, Airtable continues to operate independently.

Businesses do not need to stop using Airtable AI because of the deal. They should, however, understand exactly how AI is being used, who can access it, which models are allowed, and what data is being processed.

Why Airtable AI Governance Matters

AI inside Airtable can work directly with operational data.

Teams can use it to summarize information, analyze documents, generate content, research information, classify records, and automate repetitive tasks. Airtable currently supports models from providers including OpenAI, Google, Anthropic, Meta and others.

That flexibility is useful, but enterprises need more than access to AI features.

They need rules around:

· Who can use AI

· Which AI providers are approved

· What information AI can access

· Which workflows can run automatically

· How sensitive information is handled

· Who owns AI-generated outputs

· How AI usage is reviewed

Good enterprise AI governance is not about blocking AI. It is about making sure employees can use it without creating unnecessary security, compliance, or operational risk.

Review Who Can Use Airtable AI

The first question is simple: who currently has access?

Airtable provides centralized AI settings for Business and Enterprise Scale organizations. Admins can switch Airtable AI on or off for organization-owned workspaces and control which supported AI platforms are available.

This means enterprise teams should avoid treating AI access as an all-or-nothing decision.

Review whether:

· Every workspace actually needs AI enabled

· Users understand when AI should and should not be used

· Sensitive workflows have appropriate permissions

· AI access matches employee roles

· Former employees and external collaborators still have unnecessary access

Your Airtable AI admin controls should reflect how the organization actually operates rather than simply leaving every feature available by default.

Review Which AI Models Are Allowed

Different teams may prefer different AI models, but enterprise governance should determine which providers are appropriate for company data.

Airtable allows admins to control the AI platforms permitted within their organization. Its platform currently supports models from major providers including OpenAI, Gemini, Anthropic and Llama.

IT and security teams should understand:

· Which models are currently enabled

· Why each model is required

· Which business processes use each model

· Whether particular teams need access to every provider

· How model availability is reviewed over time

Allowing multiple models can improve flexibility, but more options also mean more governance decisions.

Understand How Airtable Handles AI Data

Airtable AI data privacy should be part of every enterprise review.

Airtable states that customer data is not used to train generative AI models, and its AI terms prohibit Airtable and third-party AI providers from using customer input or output to train generative models.

There is, however, an important difference between plan types.

For Business and Enterprise Scale workspaces, Airtable says third-party AI providers do not log input or output for human review or retain that input and output, apart from limited metadata used for billing, safety and compliance. For certain lower-tier plans, AI providers may retain input and output for up to 30 days for safety and compliance moderation.

Enterprise teams should therefore confirm which plans their AI-enabled workspaces are using rather than assuming every Airtable workspace follows identical data-retention rules.

Check Airtable AI Permissions Carefully

AI features do not automatically gain unrestricted access to an entire Airtable environment.

Airtable states that AI respects existing base and record permissions. In other words, AI can access information that the requesting collaborator already has permission to view.

That makes existing permission design an important part of Airtable AI security.

There is also an important limitation: Airtable currently does not provide field-level exclusion specifically for AI processing. Its guidance recommends using table-level and base-level permissions when sensitive fields should not be available to certain users and therefore to their AI requests.

Before scaling AI use, review whether sensitive financial, HR, customer, legal, or confidential information is properly separated.

Weak permissions become a bigger problem when AI makes it easier to search, summarize, and process large amounts of information quickly.

Know Who Controls the AI Environment

Enterprise AI governance also requires clear administrative ownership.

Airtable supports several enterprise admin roles, including Super Admin, Org Unit Admin, User Admin and Integration Admin. Super Admins can manage organization-wide security policies, SSO, SCIM, users, workspaces and other controls.

Organizations should decide who is responsible for:

· Approving AI providers

· Managing AI settings

· Reviewing AI-enabled workflows

· Investigating inappropriate AI usage

· Reviewing permissions

· Monitoring new AI functionality

· Coordinating with security and legal teams

AI governance should not depend entirely on whoever happens to build the workflow.

Audit Existing Airtable AI Use Cases

Before creating more AI workflows, understand what already exists.

An Airtable AI audit should identify where AI is being used across bases, apps, automations, and Field Agents.

For each use case, document:

· What the AI is being asked to do

· What data it can access

· Which model it uses

· Whether output is reviewed by a person

· Whether the output triggers another workflow

· Who owns the process

· What happens if the AI produces an incorrect result

Risk varies significantly by use case.

Using AI to summarize meeting notes is different from allowing an agent to update customer records, approve operational actions, or trigger downstream workflows.

The more authority an AI workflow has, the stronger the governance around it should be.

Set Rules for Human Review

Not every AI-generated result should be treated as final.

Enterprise teams should decide which workflows can operate automatically and which require human approval.

For example, AI may be suitable for:

· Categorizing incoming requests

· Summarizing documents

· Drafting content

· Researching information

· Extracting structured data

Higher-risk workflows may require review before they:

· Change financial information

· Send customer communications

· Modify important records

· Make compliance-related decisions

· Trigger external systems

The objective is not to add approval steps everywhere. It is to match oversight to the potential business impact of an error.

What Should Enterprises Monitor After the Acquisition?

The Bending Spoons deal does not currently change Airtable’s AI governance controls.

After the acquisition closes, however, enterprise teams should monitor confirmed changes involving:

· Available AI models

· AI administration controls

· Data-retention terms

· AI providers and subprocessors

· AI pricing and credits

· Permissions

· Enterprise security features

· New agent capabilities

· AI-related contractual terms

Airtable currently emphasizes that customer data is not used to train models and provides enterprise controls for AI, access management, DLP, audit logs and other security functions.

Recording the current position gives businesses a baseline against which future changes can be assessed.

Conclusion

The growth of AI inside Airtable makes governance more important regardless of the Bending Spoons acquisition.

Enterprise teams should know which AI features are being used, which models are approved, what information those tools can access, and who is responsible for managing them.

Strong Airtable AI governance means combining useful AI capabilities with clear permissions, approved models, appropriate human review, documented ownership, and a good understanding of how data is handled.

The acquisition does not require companies to pull back from Airtable AI.

It does make this a sensible time to review how AI is already being used—and make sure governance develops at the same pace as adoption.