For Rohit Khanna, chief customer officer at Smarsh, the journey into artificial intelligence wasn’t just a corporate mandate. It was a mission to prove that the technology could elevate people rather than replace them.
Smarsh operates in the high-stakes world of financial compliance, managing massive archives for 98% of the world’s largest banks. When the board pushed for an AI-first culture, Rohit turned to Salesforce Agentforce to balance cutting-edge efficiency with a human touch.
Meet the new team members: Archie and Emma
The introduction of AI initially met with significant “inertia and resistance” within global support centers. Employees feared that automation meant the end of their roles.
However, by piloting Agentforce, the narrative shifted from replacement to augmentation. Rohit and his team demonstrated that Agentforce could handle repetitive level one tickets, finally allowing the human team to step into higher-value, business-critical roles.
“It will make our level one support more efficient and we can actually promote level one to level two to level three,” Rohit explains, noting that for 2026, the company is planning to promote 25% of its level one staff.
At Smarsh, Agentforce isn’t just a tool; it’s a digital workforce with a personality. The company deployed two distinct agents via the Agentforce platform:
- Archie: The external-facing agent who serves as the first point of contact for 22,000 users on the Smarsh Central portal.
- Emma: The internal agent built on Agentforce to help employees summarize data and create knowledge base articles.
The impact of Agentforce has been “phenomenal.” Archie is currently deflecting 68% of incoming calls, enabling humans to focus on complex, nuanced problems that require real expertise.
Three steps to Smarsh’s Agentforce success
Rohit is the first to admit that this success required a disciplined approach. To move from 45% to 86% accuracy over six months, Smarsh followed a pragmatic framework:
- Clean the Data: Smarsh spent years on data hygiene before the first bot went live. Rohit warns: “Do not buy any agentic AI technology… unless you are very sure that your data is good.”
- Govern the Data: Especially in the financial sector, strict governance ensures that Agentforce “should not, could not, will not” render data to unauthorized users.
- Include a Human in the Loop: Smarsh employs a team of nine people to supervise Archie, ensuring the agent is “adult enough” to represent the brand without hallucinating.
The future: agentic workflows
The next frontier for Smarsh is the Agentic Workflow. Within the next six to nine months, the company expects Emma and Archie to talk to each other — passing assignments back and forth and learning from one another.
By leading with Agentforce, Smarsh improved its bottom line and built a smarter, more fulfilled workforce. For Rohit Khanna and his team, Agentforce isn’t just a software implementation — it’s a journey toward a more productive future.
Learn more considerations and best practices in deploying Salesforce Agentforce.
Meet Agentforce for Service
Watch Agentforce for Service resolve cases on its own, deliver trusted answers, engage with customers across channels and seamlessly hand off to human service reps.
