Legal
Sub-processors
Last updated: 29 September 2026
What this page is
A sub-processor is a third party that processes customer data on our behalf so that we can deliver our services. This page names the ones we use, what each does, the data involved, and where it is processed. It is the list referred to in our Data Processing Agreement.
It covers only vendors that process customer personal data for us. Tools we use to run our own business — our marketing and sales systems, for example — are not listed here, because the data in them is ours rather than processed on a customer's behalf.
Our sub-processors
| Sub-processor | Purpose | Data Processed | Location |
|---|---|---|---|
| Amazon Web Services, Inc. | Cloud hosting and infrastructure — compute, database, object storage, transactional email | All Subscriber Data | United States (us-east-1; backup copies in us-east-2) |
| Microsoft Corporation | Identity services (Azure Entra ID); dataset hosting and analytics serving (Power BI) | Membership and engagement data held in per-Subscriber datasets | United States (East US 2) |
| Google LLC | Two distinct purposes. (1) Analytics export: Company operates per-Subscriber Google Cloud projects that export Google Analytics 4 data for Subscriber’s own web properties into BigQuery. (2) Identity services: single sign-on, where Subscriber elects to authenticate its Authorized Users via Google Workspace | (1) Analytics event data relating to visitors to Subscriber’s web properties: pseudonymous identifiers, IP-derived approximate location, device and browser characteristics, and behavioural events. (2) Authentication data only: user identifier, email address and directory attributes returned at sign-on | United States |
| Anthropic, PBC | AI Features — large language model inference across the in-product assistant, scheduled insight generation, recurring prompts, programmatic access, file inspection, public chart chat, and internal review of assistant conversations | Subscriber Data submitted to or Processed by AI Features. See Sections 7.1 and 7.3 | United States |
| OpenAI, L.L.C. | AI Features — large language model inference across the same surfaces, for Subscribers whose configured AI provider is OpenAI. Each Subscriber is configured to a single provider, so Subscriber Data reaches one large language model provider and not both | Subscriber Data submitted to or Processed by AI Features. See Sections 7.1 and 7.3 | United States |
| Slack Technologies, LLC | Internal operational monitoring, account management and customer support — every turn of every assistant conversation, across all tenants, is mirrored in real time to a staffed internal channel (#kraken-chat-firehose). A single global mirror, enabled 20 May 2026 | The authenticated user’s email address and the Subscriber’s organization name are transmitted in full, as named fields in every message. Conversation free text (the prompt and the response) is scrubbed of Personal Data before transmission; that scrub is automated, designed to be thorough, and not represented as exhaustive. Also transmitted: a truncated session identifier, token counts, duration and outcome | United States |
Changes to this list
We may engage additional sub-processors, or replace one with another, within the purposes described above. When we do, we update this page and give notice in accordance with our Data Processing Agreement, which sets the notice period and the customer's right to object.
Where a sub-processor changes something on its own side and we only learn of it once it has taken effect, we tell customers promptly after we find out rather than in advance, because advance notice is not ours to give in that situation.
Questions
Write to us at A2 Holdco, LLC, DBA Association Analytics, Attn: Data Privacy, 405 E Laburnum Ave Ste 3, Richmond, VA 23222, or see our Privacy Policy.