Retail Distributor Scales Live Chat with Thunai in Salesforce

Client Overview

Our client is a Major North American Retail Distributor catering to facility managers, property owners, and purchasing professionals across the USA.

Their support team at our client's company receives numerous live chat sessions daily. These typically range from customer service inquiries about order statuses and products to billing concerns and account-related queries.

With Salesforce at the center of their support operations, the client needed a way for agents to keep up with live chat demand. And do this without lowering the quality of responses, dropping under pressure.

The Challenges: Slow Replies, Uneven Quality, Disconnected Knowledge

Live chat puts a different kind of pressure on support teams. Customers expect fast replies in real time, and agents were struggling to keep up without a reliable system behind them. Three problems kept surfacing across the team day after day:

• Inconsistent Customer Experience: With too many chats happening simultaneously, customers received different answers depending on which agent handled the conversation. Doing this caused a delay in their response time, thus creating backlogs. Customers waited more than necessary for answers to questions that had been previously handled several times.
  
• Growing Knowledge Dependency: Without any common benchmark, agents relied on their personal experiences in drafting their reply. As product information expanded, agents struggled to consistently apply policies and procedures. But also, his/her recall ability of procedures followed in the company, and not on any common benchmark set by all the agents collectively.
  
• Limited Knowledge Reuse: Successful resolutions remained trapped within individual conversations rather than benefiting the wider support organization. This meant that the next agent dealing with such queries the next day had no idea how the previous agent had solved the problem, nor the results of his efforts.

Our Intelligent Solution

We placed Thunai directly inside the client's Salesforce environment as a native LWC widget. Agents stayed in the same interface. Thunai sat alongside every live chat window and worked in step with the conversation as it played out.

• Real-Time Chat Reply Suggestions: As a customer message comes in, Thunai reads the full conversation context and surfaces a suggested reply directly within the Salesforce interface. Agents review it, make any adjustments, and send. In doing this, the time between a customer sending a message and an agent getting a reply drops significantly, without any drop in response quality.
  
• Documentation-Backed Replies via Thunai Brain: The client's SOPs, product guides, warranty policies, and historical ticket data were all fed into Thunai Brain. Every reply suggestion Thunai puts forward is drawn from this content, so agents are not working from memory or making judgment calls on policy. With this, customers consistently get answers that match actual company guidelines rather than individual interpretations.
  
• Agent-Directed Reply Instructions: Agents can tell Thunai directly to modify the policy details with any corrections and the key points to address. Approved guidance and corrections that can be incorporated into future recommendations through Thunai's knowledge management workflow.
  
• Live Case Sync via Thunai MCP: Every chat interaction, suggested reply, and agent-made adjustment writes back into the Salesforce case record automatically as the conversation closes out. Every interaction and agent action is automatically written back into Salesforce, providing supervisors with complete auditability, reporting visibility, and quality assurance insights

Conclusion

Thunai gave the support team a way to keep up with live chat volume without asking agents to work faster on their own or cut corners on the replies they were sending out.

By combining real-time AI guidance with documentation-backed responses, Thunai helped the client improve response consistency, reduce handling times, and scale support operations without increasing agent workload.

The support team now spends less time searching for answers mid-chat and more time closing conversations out to a resolution.