Smarsh Scales Salesforce's Agentforce Following 72% Support Deflection Success

Smarsh has released performance data for its AI-powered customer support agents, Archie and Emmy, built on Salesforce’s Agentforce platform. For fintech leaders, these results provide a rare benchmark for "agentic AI" in highly regulated environments, demonstrating that automated systems can handle complex compliance requirements while significantly reducing human workloads and improving resolution speeds.

What was announced

Smarsh revealed that its customer-facing AI agent, Archie, achieved a 72% deflection rate during the second quarter of 2026. Out of 405 interactions recorded in that period, more than seven in ten sessions were resolved without the need for human intervention. The agent also secured a 2.6 out of 3 resolution confidence score, a metric Smarsh uses to measure whether a response fully resolved a query based on manual review and the absence of customer escalation within 48 hours. A score of 3 indicates high confidence, including no frustration or challenge to the response during the conversation.

The company also detailed the performance of Emmy, an internal AI agent launched in late March 2026 to assist support representatives with account snapshots and resolution guidance. Emmy reached a 65% adoption rate among staff by the end of June, with 120 out of 185 representatives utilizing the tool. According to the data, 31% of cases have been resolved using AI since Emmy's launch, surpassing the company’s initial 20% target. Notably, teams using Emmy saved an average of 7.5 hours per case compared to similar cases handled by the same staff before the AI's introduction. This efficiency gain exceeded the company's internal target of five hours per case, particularly in scenarios involving system-log analysis and engineering escalations. Phil Dean at Smarsh will present further lessons from this implementation at Dreamforce on September 16.

"A year ago, we set out to use AI in a way that would create tangible value for our customers and our teams. The most important result isn’t simply that Archie can resolve routine needs faster. It’s that our customers can trust us to get a consistent experience while our people have more time to focus on the complex issues where their knowledge and judgment matter most. Those results give us a model we can now scale responsibly across the business."

Rohit Khanna, Chief Customer Officer at Smarsh.

The companies involved

Smarsh, headquartered in Portland, Oregon, is a specialist in communications data and intelligence. The firm provides services to regulated organizations, including many of the world's largest banks, insurers, and government agencies, helping them capture and analyze digital communications to manage risk and maintain compliance. The company has recently secured a Notice of Allowance from the U.S. Patent and Trademark Office for the Archie trademark, signaling a move to formalize its AI branding as it moves from initial experimentation to enterprise-scale deployment.

Salesforce, the provider of the Agentforce platform used by Smarsh, is a global leader in customer relationship management (CRM) software. Agentforce represents the company’s push into agentic AI, designed to embed autonomous capabilities directly into enterprise workflows. The partnership between the two firms highlights a growing trend of integrating specialized compliance knowledge with broad-scale AI infrastructure. Greg Beltzer, Chief Customer Officer, SVP-Agentforce at Salesforce, noted that Smarsh serves as an example of how enterprises in regulated sectors like financial services can scale AI while meeting strict governance requirements demanded by regulators.

What FF News has reported before

FF News has frequently tracked the expansion of Salesforce’s AI ecosystem within the financial services and insurance sectors. Recently, we reported on how Sixfold and PS Advisory Launch AI Underwriting Integration for Salesforce via MuleSoft, demonstrating the platform's utility in complex underwriting tasks. The movement of high-level talent also underscores the platform's influence; FF News covered when Waystar Appoints Former Salesforce AI Leader Amit Khanna as CPTO to Drive Autonomous Revenue Cycle. These developments reflect a broader industry shift toward autonomous operations, though consumer sentiment remains cautious, as noted in our report where NMI Research Reveals 70% of Consumers Demand Final Say Over AI-Driven Purchases.

What this means

The results from Smarsh suggest that the industry is moving past the "experimental" phase of AI chatbots and into a period of measurable operational impact. The 7.5-hour time saving per case for Emmy is particularly significant; it suggests that AI’s greatest value in fintech may lie in internal "copilot" roles that handle the heavy lifting of data analysis rather than just customer-facing chat. This puts pressure on traditional customer service outsourcing models that rely on headcount rather than efficiency. The primary question now facing the sector is whether these efficiency gains will lead to a reduction in total workforce or if the saved hours will truly be reinvested into complex troubleshooting as the company suggests.