Thunai Improves Salesforce Email Ticket Closure 3X for a Fortune 100 Retailer

Client Overview

Our client is a Fortune 100 company operating as a retailer and distributor for facility managers, property managers, and procurement organizations in the U.S.

Their customer support team manages thousands of incoming emails every day, including issues related to orders, warranty on products, disputes about billing, and tracking deliveries.

Since Salesforce was the platform for all their customer support services, they needed a solution that would help them manage the flood of incoming emails.

And do this, without compromising the quality of responses or exhausting customer support teams that need to type emails manually.

The Challenges: Slow Replies, Inconsistent Responses, No Central Knowledge

Agents were spending a large part of every working day reading through email tickets and writing replies from the ground up. But with this, three problems kept building up across the team:

• Replies Written From Scratch Every Time: Every incoming email ticket required an agent to draft a full response independently, pulling from memory or digging through internal documents to piece together the right answer. With this, resolution times ran long and the quality of responses depended heavily on the individual agent handling the ticket.
  
• No Single Source for Policies and Procedures: SOPs, warranty guidelines, and escalation procedures were spread across separate documents with no central place to pull from. Agents frequently went back and forth with supervisors before sending a reply, slowing the whole queue down. Meaning, tickets that could have been closed in one response were often taking several exchanges to wrap up.
  
• Replies That Did Not Carry Over Learning: When an agent worked out the right way to address a particular type of issue, that knowledge stayed with them. The next agent picking up a similar ticket started from zero again, with no record of how the previous resolution was worded or reasoned through.

Our Intelligent Solution

We placed Thunai directly inside the client's Salesforce environment as a native LWC widget. Agents did not switch to a new platform. Thunai sat within the interface they already used every day.

• Email Response Suggestions within Salesforce: Thunai reads through every email ticket received and recommends an answer right in the Salesforce platform. Agents look through the suggested answer, tweak information or policy details if needed, and then send it out. The result was that the time spent per ticket decreased since the process shifted from writing to proofreading.
  
• Policy-Accurate Replies via Thunai Brain: The client's SOPs, warranty policies, and internal procedures were fed directly into Thunai Brain. Thunai draws from this content when generating every reply suggestion, so responses are grounded in actual company documentation rather than individual agent recall. But on the other hand, where a ticket falls outside standard policy, the system flags this clearly so the agent can take it forward with full awareness.
  
• Agent-Directed Reply Instructions: Agents can give Thunai direct instructions on how a particular type of email should be addressed, down to tone, structure, and the specific details to call out. With this, Thunai carries those instructions forward and applies them automatically to all future tickets of a similar nature, so the team does not have to keep spelling out the same guidance repeatedly.
  
• Full Ticket History Sync via Thunai MCP: Every suggested reply, agent adjustment, and closed ticket writes back into the Salesforce case record automatically. Team leads can track what was sent, how tickets were handled, and where the queue stands at any point without chasing agents for updates.

Conclusion

Thunai closed the gap between an agent opening a ticket and getting a reply out the door, without asking the team to change how they work or where they work.

Reply suggestions grounded in actual company documentation mean agents are no longer piecing together responses from scattered sources. In doing this, the back-and-forth that used to slow down straightforward tickets has largely dropped away, and the queue moves faster as a result.

The instruction-retention feature means the team's collective knowledge builds up over time rather than staying locked with individual agents. 

Meaning, a resolution pattern that one agent works out on a Monday is already informing replies across the team by the end of the week. The support team now spends less time writing and more time closing tickets.