Smarsh unveils AI agents to speed probes & cut costs

Smarsh has launched a set of AI-based software agents designed to speed up corporate investigations and reduce the cost of legal discovery and compliance surveillance.

Its new Discovery Agent aims to shorten investigation timelines from weeks to hours and cut discovery costs for legal teams by up to 75% by surfacing relevant information earlier in a matter.

Smarsh has also released an Intelligent Agent for compliance surveillance, which it says can help teams spot risk signals sooner and reduce the number of low-risk alerts that require review.

Discovery focus

Legal discovery remains a significant operational cost for large regulated organisations, especially as employees and clients use a growing range of communications channels. The rise of AI-generated content has also increased the volume of material legal and compliance teams may need to assess during investigations and litigation.

According to Smarsh, the Discovery Agent uses AI to highlight information likely to matter early in an investigation, improving early case assessment and reducing reliance on outside counsel.

Smarsh describes the Discovery Agent as part of a "unified workflow" that combines several investigation steps, including reconstructing timelines, identifying key custodians, and visualising communication patterns across multiple languages.

Kim Crawford Goodman, Chief Executive Officer of Smarsh, said the launch reflects a shift in how organisations treat archived communications.

"Compliance is entering a new era where communications data can no longer sit idly in the archives," said Kim Crawford Goodman, CEO of Smarsh. "Our customers are transforming their data into an actionable intelligence engine that accelerates investigations and drives faster decisions. We're proud to stand at the forefront of this shift, providing the tools that transform regulatory burden into a strategic advantage."

Surveillance changes

The Intelligent Agent is designed for day-to-day monitoring of employee communications in regulated environments. Smarsh says the product is now "in full production" and uses AI filtering to reduce the volume of alerts compliance teams must triage.

The company says contextual filtering suppresses low-risk alerts and can cut review volumes by up to 50%. It also says its multilingual "Detect Scenarios" can identify three to five times more real risks, helping address alert fatigue in financial institutions.

Smarsh says the agents run on domain-adapted large language models and that the platform includes transparency and auditability features aligned with regulator expectations, though it did not name specific supervisory regimes.

Data and storage

The launch also includes infrastructure updates focused on how communications records are stored and made available for analysis, including enhanced security and faster API integrations.

As part of the update, Smarsh introduced CryoStore, a new storage tier it says supports the "full data lifecycle" by keeping years of historical records available for compliance while allowing organisations to synchronise AI-ready data.

Smarsh is pitching CryoStore as a lower-cost option for long-term retention, saying it is available at "a fraction of the traditional cost," without providing pricing details.

Goutam Nadella, Chief Strategy Officer of Smarsh, said data readiness is central to the performance of AI models used in compliance and risk work.

"AI is only as powerful as the data behind it," said Goutam Nadella, Chief Strategy Officer at Smarsh. "When communications data is fragmented or inaccessible, model performance plateaus. By prioritizing data readiness today, with solutions like CryoStore, organizations can fully unlock the full value of AI for risk management tomorrow."

Market context

Regulated firms have invested heavily in communications capture, archiving, supervision, and e-discovery over the past decade, tracking the spread of messaging platforms beyond email and regulators' continued focus on recordkeeping and supervision.

Smarsh is positioning its agent-based approach as a move away from manual review and post-incident investigations. The company says it aims to shift oversight from reactive monitoring to proactive intelligence, focusing on earlier signal detection and lower operational spend.

It says the new agents are intended for legal and compliance teams handling growing volumes of communications data and more complex investigations across languages and channels.