Airtable added Anthropic and dropped an AI-agent vendor

The change

On July 6, 2026, Airtable updated its Subprocessors page. Venpo detected the change and separated the substantive edits from the formatting noise.

Inside a routine-looking refresh sit four real changes. Anthropic was added as an AI model provider. Sampark was removed from the AI-agent workflow. Castle Global's AI content moderation was narrowed to Enterprise and Business plans. And Pylon Labs was added as a customer support subprocessor.

Airtable sits in the operations and data layer of a large number of companies, and its Airtable AI features route customer data through this exact list of providers. A change to who is on it reaches everyone using those features.

What changed

The update touched three parts of the list.

In the In-Product AI section:

• Anthropic, PBC was added as a US-based AI model provider

• Sampark was removed, along with its role as the integration authentication and execution layer for Omni and AI agents

• Castle Global's AI content moderation was marked as applying only to Enterprise and Business plan customers

In the Other Third-Party section, Pylon Labs was added as a US-based customer support subprocessor.

The rest of the diff was cosmetic: punctuation cleanups on company name suffixes, a reformatted table header, and a few rows that moved position without changing meaning.

The AI vendor set keeps shifting

A single subprocessor edit is rarely a headline. What makes this one worth attention is the direction of travel. With Anthropic added, Airtable AI now lists five separate AI model providers: Amazon Web Services, Anthropic, Google, IBM, and OpenAI. Alongside them sit specialized AI vendors for content moderation, internet search and browsing, data extraction, and web prototype generation.

The set does not only grow. In the same update, Sampark left the list and Castle was narrowed to higher plan tiers. That is the point: the roster of external AI vendors that can process Airtable customer data is edited in both directions, on Airtable's own schedule, not on any customer review cycle.

For a company that approved Airtable six months ago, the AI providers in the chain today are not the ones that were vetted at onboarding.

Why this matters

Subprocessor lists are one of the few places where a vendor discloses exactly who else touches your data. They are also among the most frequently edited documents a vendor maintains, and changes are almost never announced beyond a shifted revision date. This is the exact category of change covered in our analysis of the hidden risk of vendor legal changes.

When AI providers are added and swapped at this pace, two things happen. The data exposure surface moves, and the documentation that security and compliance teams approved at onboarding falls out of date. The same dynamic is visible across the stack, from Anthropic narrowing its data retention commitments to RingCentral adding AI subprocessors almost every month. AI vendors are being woven into vendor stacks faster than governance processes were built to handle.

Potential impact for SaaS companies

Companies using Airtable AI may want to review whether:

• their data processing records still match Airtable's current subprocessor list, including Anthropic and Pylon Labs

• their security and compliance approval accounted for customer data being processed by an additional AI model provider

• customer-facing privacy documentation and DPAs reflect the AI providers now in the chain

• the plan-based limit on Castle changes anything for their tier

For teams that maintain their own subprocessor disclosures downstream, a change like this can cascade: your vendor's new subprocessor may need to appear in your own documentation. Tracking that across a full portfolio is where monitoring AI subprocessors becomes a structural necessity rather than a periodic task.

How Venpo detected it

Venpo continuously monitors vendor legal documents and detects changes as they are published. When Airtable updated its subprocessor list, Venpo:

• detected the update and the shifted revision date

• separated the four substantive changes from the punctuation and reflow noise

• identified Anthropic and Pylon Labs as additions, and Sampark as a removal

• flagged the plan-based narrowing of Castle as a separate, positive nuance

The full redline of the change is public on our live feed: Airtable Subprocessor List, July 6, 2026. Building continuous vendor risk monitoring into the compliance stack is what makes this kind of detection possible.

Business outcome

Companies that caught this change early were able to:

• update their vendor and data processing records to include Anthropic and Pylon Labs

• review whether an additional AI model provider fits their internal data handling policies

• confirm their customer-facing disclosures still match reality

• brief security and compliance before the next audit or customer review

Instead of discovering a new AI vendor in the chain during due diligence or a customer security questionnaire, they adapted on their own schedule.

Key takeaway

Subprocessor lists are where AI sprawl shows up first, and they change faster than almost any other vendor document, in both directions. A single update can add a major AI model provider, remove another vendor, and narrow a third by plan tier, all at once. The only way to know it happened is to be watching the document when it changes. A closer look at why scheduled reviews keep missing this kind of change is in our comparison of manual vs automated vendor monitoring.