4 insights from Smarsh EDGE: Governance becomes the foundation for communications intelligence

At Smarsh EDGE in New York, the old compliance archive got a promotion. Employee messages, once treated mainly as records for regulators, are now being framed as a source of business intelligence, with AI able to sift through a year of email, chat and prompt traffic and turn it into usable context.

That sounds tidy until you remember AI is now part of the communications stack too. Financial firms are already building their own agents outside compliance platforms, then pulling that data through Model Context Protocol servers. Smarsh’s point is blunt: regulations do not care that the software is new, and firms still carry the liability when those agents act.

Gemma Allen of theCUBE put the data angle plainly, calling it the battleground where system of record becomes system of engagement and then system of intent. Smarsh says its own study found only 26% of enterprises believe their governance frameworks are keeping pace with AI deployment, even as financial services moves past pilots. That gap is the real story here.

The event’s interviews kept circling back to the same idea from different angles. Goutam Nadella said AI can classify conversations across email, WhatsApp, Teams and Zoom, while Rohit Khanna showed how supervisors can ask for plain-language answers instead of wrestling with Boolean searches. Kamesh Tumsi described four levels of autonomy, from assisting to fully autonomous, and said each workflow needs to be matched to its risk, reversibility and observability. Meanwhile, Smarsh is leaning on a multiyear collaboration with AWS and work with Caylent to push governance deeper into the infrastructure layer.

The throughline is simple: the winners will not be the firms that use more AI, but the ones that know exactly what it touched, who touched it and why. In regulated markets, that is less sexy than a chatbot and much more valuable.