SOC 2
SOC 2 is an audit report, produced by an outside CPA audit firm under AICPA standards, on a service organisation's controls relevant to security, availability, processing integrity, confidentiality or privacy.
Also called: SOC 2 Type II, SOC 2 report, Trust Services Criteria
By Divya Sharma · Checked Nov 15, 2026 · How we review
When a company buys an AI tool or API, its security team often asks for the vendor's SOC 2 report. The report is written by an outside auditor and describes the vendor's systems, the controls it has in place, and how those controls were tested. It is defined by the American Institute of Certified Public Accountants (AICPA).
What it covers
The controls are measured against the AICPA Trust Services Criteria: security, availability, processing integrity, confidentiality and privacy. A vendor chooses which criteria are in scope, so two SOC 2 reports can cover quite different things. A Type II report, the kind large cloud providers publish, tests both the design and the operating effectiveness of controls across an audit period, not on a single day.
How you get one
SOC 2 reports are usually shared under a non-disclosure agreement. AWS, for example, requires an NDA to review its SOC 1 and SOC 2 reports, and offers a SOC 3 report as a public summary.
What it does not tell you
A SOC 2 report is an attestation about controls, not a guarantee that a breach cannot happen, and it covers only the systems in its scope. For AI vendors, questions such as whether prompts are stored or where they are processed sit alongside it; see zero data retention and data residency. Rules specific to AI systems in Europe are set by the EU AI Act. This page describes the term; it is not compliance advice.
Related terms
- Data residency
- Data residency is a control that keeps a customer's data stored, and often processed, in a chosen geographic region such as the EU, so a team can meet its own legal, contractual or customer rules on where data lives.
- Zero data retention
- Zero data retention (ZDR) is an arrangement in which an AI provider does not keep a customer's prompts and outputs after a request is answered, usually offered only to approved customers and only for eligible API features.
- EU AI Act
- The EU AI Act, Regulation (EU) 2024/1689, is the European Union's law on artificial intelligence. It sets risk-based rules for those who build and deploy AI systems, from outright bans to transparency duties, phased in from 2025.
- Inference provider
- An inference provider is a company that runs AI models on its own hardware and sells access to them through an API, usually charged per token. You send prompts and get outputs without managing GPUs yourself.