Zero data retention is one of the most requested terms in AI vendor contracts, and one of the most misunderstood. It is a precise commitment about one layer of the stack, not a promise that nothing is stored anywhere.
This guide explains what zero data retention means, where it applies, what it does not cover, and the questions to ask before relying on it.
What Is Zero Data Retention?
Zero data retention is a policy under which an AI provider does not store the inputs sent to its models or the outputs they return once a request has been processed. Under zero data retention, prompts, transcripts, and generated text are handled in memory and discarded, so they are not kept in logs, not reviewable later, and not used to train models.
Retention Options Compared
| Standard retention | Limited retention | Zero data retention | |
|---|---|---|---|
| What is stored | Inputs and outputs, often for 30 days or more | Inputs and outputs for a short, fixed period | Nothing after the request completes |
| Why | Abuse monitoring, debugging, support | Abuse monitoring with reduced exposure | Minimum exposure |
| Model training | Depends on the contract | Usually excluded by contract | Excluded |
| Typical availability | Default consumer and basic plans | Business plans | Enterprise agreements, often after eligibility review |
How Does Zero Data Retention Work?
Zero data retention is a combination of engineering and contract.
- Request. The application sends a prompt, such as a transcript, to the model provider.
- Processing in memory. The model generates the response without writing the request to persistent storage.
- Response. The output returns to the application.
- Discard. Nothing about the request is retained by the model provider.
- Contract. The commitment is written into the agreement, with any exceptions stated, such as safety monitoring with very short retention.
- The application layer is separate. The vendor whose product you use may still store transcripts, summaries, and records in its own systems under its own policy.
That last point is the one most buyers miss. Zero data retention at the model provider says nothing about what the application vendor keeps.
Key Features of Zero Data Retention
- No persistent storage of inputs or outputs. Data exists only during processing.
- No model training. Customer data never improves the provider’s models.
- Contractual commitment. The policy is in the agreement, not only on a web page.
- Stated exceptions. Any safety or legal exceptions are explicit and narrow.
- Applies at a named layer. The commitment identifies which party and which systems it covers.
Benefits of Zero Data Retention
- Smaller breach surface. Data that is not stored cannot leak later.
- Simpler compliance. Fewer copies of personal data to govern.
- Customer confidence. Buyers in regulated industries often require it.
- Faster procurement. A clear answer to a common security question.
Zero Data Retention Use Cases
- Sales conversation AI. Transcripts contain personal and commercial data, so model-layer retention matters.
- Regulated industries. Financial services, healthcare, and legal buyers often require it.
- Cross-border data. Less stored data means fewer transfer questions.
- Vendor evaluation. Zero data retention is one line in a broader review of the vendor’s own storage, access, and retention.
How ASPR fits. ASPR processes sales conversations and CRM data, so data handling is part of every customer review. ASPR publishes its security and data practices on its Security page and Trust Center, which is where to confirm the specific commitments that apply.
See ASPR's agents on your own deals
Watch AI agents prepare your reps, capture every conversation, update your CRM, and coach your team.
Book a DemoFrequently Asked Questions About Zero Data Retention
Related terms and internal links
- Related glossary terms:AI Data Privacy,Enterprise AI Security,Security & Compliance,SOC 2 Compliance,Enterprise AI Governance,Sales Call Recording
- Existing pages:ASPR Security,ASPR Trust Center
Sources
- GDPR, Regulation (EU) 2016/679, Article 5 on storage limitation
- AICPA, SOC 2 reporting and the Trust Services Criteria (aicpa-cima.com)