.jpg)
In today's enterprise AI landscape, data governance has become a strategic pillar. With the arrival of foundation models that require sharing information with external providers, organizations need robust mechanisms to control data retention without losing the ability to innovate. Amazon Bedrock offers several configuration levels, from account-level retention modes to service control policies (SCPs) that operate across the entire organization. However, the key is understanding that the configured mode acts as a ceiling, not a floor: it allows using models that require less retention, but does not force retaining more than necessary.
This permission architecture is especially useful when combining isolated projects within the same account using Amazon Bedrock Projects, which are only available on the bedrock-mantle endpoint (compatible with OpenAI and Anthropic APIs). Thus, a research team can enable the provider_data_share mode to access the latest models, while a production application handling sensitive data remains in none mode. For environments using the bedrock-runtime endpoint, segmentation is achieved through separate accounts in AWS Organizations, grouped into organizational units (OUs) with different policies.
Using SCPs is the most drastic and effective measure to guarantee zero retention across the entire organization. By attaching an SCP to the root OU, any attempt to change the retention mode to values other than none is blocked, including actions on projects in bedrock-mantle. It is important to note that AWS Organizations management accounts do not inherit these restrictions, so complementing with IAM policies is recommended. Verification via AWS CLI or the Bedrock API allows confirming that the policy is active.
For companies looking to implement these strategies securely and efficiently, having a specialized technology partner makes the difference. At Q2BSTUDIO, we offer AWS and Azure cloud services that integrate data governance, artificial intelligence, and cybersecurity architectures from the design phase. Our team develops AI solutions for businesses that meet the strictest compliance requirements, whether through custom applications, AI agents, or business intelligence services like Power BI. The combination of custom software with granular retention policies allows our clients to adopt cutting-edge models without compromising their data privacy.
Ultimately, zero retention is not a limitation, but a strategic decision that requires tools, policies, and expert guidance. With the correct implementation of SCPs, projects, and a technology partner like Q2BSTUDIO, any organization can navigate the Amazon Bedrock ecosystem with full control over their information.
