At the Smarsh EDGE event, the company highlighted a shift in how financial institutions manage communications, moving from simple archiving to extracting business intelligence using artificial intelligence. Smarsh is now capturing AI-prompt traffic, including agent-to-agent communications, to ensure regulatory compliance.
A study by Smarsh Inc. indicates that only 26% of enterprises believe their governance frameworks are keeping pace with AI deployment. To address this, Smarsh has partnered with Anthropic PBC and OpenAI Group PBC to provide compliance APIs that track prompts and trace risky outputs to their sources.
Smarsh has introduced new tools, including a patent-pending context layer for identity mapping to reduce false positives in risk detection. The company also developed specialized compliance language models, which are described as more cost-effective than standard large language models.
The firm has released several agents over the past six months and launched a new Model Context Protocol server. According to Smarsh, 70% to 80% of compliance spending is currently allocated to labor, suggesting significant potential for AI-driven cost reduction in the sector.
Smarsh is collaborating with Amazon Web Services Inc. through a multiyear agreement to integrate governance and security at the infrastructure layer. This partnership aims to provide customers with better visibility into data access as they move toward more autonomous AI systems.
