Compliance Becomes Strategic: How Banks Are Mining Employee Communications for Business Intelligence

Banks and financial services firms have historically viewed the preservation of worker communications as a compliance obligation—a necessary evil for regulators. But at the Smarsh EDGE event, that perspective shifted dramatically. The industry is now recognizing these vast repositories of email, instant messages and AI-generated content as untapped sources of business insight. Machine learning systems can now process months or years of communications traffic and extract actionable context. The challenge intensifies as AI itself becomes another communication channel requiring capture and oversight, including machine-to-machine agent interactions.

The coming era will be shaped by governance frameworks embedded into systems from inception rather than added as an afterthought. Many financial institutions are developing proprietary AI agents outside traditional compliance infrastructure and connecting them via Model Context Protocol servers. Yet regulations remain indifferent to the technology stack itself, leaving firms legally accountable for whatever their agents do. A recent Smarsh Inc. study revealed a troubling gap: merely 26% of organizations believe their governance structures are keeping up with the pace of AI rollouts, even as financial services firms move past experimental phases.

I think that move from system of record to system of engagement to system of intent, it's really won or lost in one battleground, and that is data. Data is a huge moat for companies in 2026.

Gemma Allen, host at theCUBE, SiliconANGLE Media's livestreaming studio

Allen moderated conversations with industry leaders during the Smarsh EDGE gathering in New York. Discussions ranged across communications intelligence, compliance-focused language models, degrees of agent autonomy and cloud infrastructure partnerships underpinning Smarsh's offerings.

Four key takeaways from the event

1. Communications intelligence converts messy message streams into relationship insights

Smarsh is enabling firms to extract business value from communications that were previously too fragmented and disorganized to leverage effectively, according to Goutam Nadella, chief strategy officer of Smarsh. Rather than relying on what a CRM system captured about a single meeting, AI can now review a full year of dialogue spanning email, WhatsApp, Teams and Zoom, then recommend preparation strategies for upcoming interactions. Smarsh participated as an early partner in compliance APIs developed by Anthropic PBC and OpenAI Group PBC, enabling the capture of prompts so problematic outputs can be traced to their origins. Nadella emphasized that governance must be woven into the core architecture from day one, not treated as an occasional check—a lesson underscored by the billions in fines levied against financial institutions for failing to properly retain business communications on platforms like WhatsApp.

2. Plain-language queries replace complex search syntax for compliance teams

Compliance officers who previously constructed intricate Boolean queries to detect market manipulation can now pose questions to an AI agent in everyday language and receive flagged exchanges complete with risk rankings, according to Rohit Khanna, chief customer officer of Smarsh. A proprietary context layer—for which Smarsh holds a patent—that understands organizational hierarchies and identity structures helps the system avoid generating spurious alerts. The firm has also developed specialized compliance language models that consume fewer resources than general-purpose large language models, including a version supporting multiple languages that can track when speakers code-switch. An internal innovation group called S-Combinator, comprising 11 volunteers, created a model risk management agent currently in limited release. Khanna reiterated that governance must be the starting point.

3. Agent-driven automation adjusts autonomy levels based on workflow risk

Firms can combine Smarsh's agents with their own models and agents while relying on the platform for underlying governance, data permissions and interpretability, noted Kamesh Tumsi, chief product officer of Smarsh. The company has rolled out four or five agents over six months alongside its new MCP server. Since 70% to 80% of compliance budgets still flow toward human labor, the upside from AI automation is substantial, according to Tumsi. He outlined four tiers of autonomy: assisting humans, recommending actions, executing within defined boundaries and operating fully autonomously. Each workflow gets assigned a tier based on whether the action is reversible, its technical complexity and the level of visibility required.

4. AWS and Caylent embed governance into the infrastructure foundation

Skepticism about AI deployment has transformed into enthusiasm now that supporting infrastructure is in place, Nadella observed during a panel that included Arym Diamond, head of data and AI go-to-market sales at Amazon Web Services Inc., and Peter Sankauskas, portfolio chief technology officer of Caylent Inc. Under a multiyear strategic partnership with AWS, Smarsh is integrating governance and security controls at the infrastructure level, responding to customer demands for visibility into who and what accesses their information. Diamond explained that Smarsh has transitioned from functioning as a system of record to operating as a system of engagement powered by AWS infrastructure. Sankauskas added that AI requires equilibrium: authority without context produces poor outputs, while context without authority forces teams to constantly supervise the system.