Using AI to improve customer experience across its brands.

Turning AI into a better customer experience.

Buying a mattress is personal. Customers want the right product, clear answers, and confidence in their purchase. When something goes wrong, they want support that’s fast, helpful, and human.

For more than 100 years, Serta Simmons Bedding has helped people find comfort through its Serta, Beautyrest, and Tuft & Needle brands. As its direct-to-consumer business grew, customer service became more complex, spanning product questions, purchases, returns, warranties, and retail partner support.

For Steven Cook, Program Manager, Customer Experience, the goal is straightforward: help customers get the answers they need, quickly.“Our goal is to meet customers where they are, and get answers to them quickly,” Cook says.

With Talkdesk Customer Experience Automation (CXA), Serta Simmons Bedding combines AI-powered self-service and agent assistance to resolve routine questions faster while preserving the human support customers need for more complex situations.

Making customer support more conversational.

As Tuft & Needle’s digital business expanded, staffing constraints forced the company to scale back live chat and direct more inquiries to email. Response times could stretch to four days, leaving customers waiting for help with everything from product questions to time-sensitive price adjustments.

The team restored chat and used live-agent conversations to understand which needs were best served through each channel. It then expanded Talkdesk Autopilot to automate common questions about products, returns, and warranties.

However, customer feedback revealed that the initial decision-tree experience, while accurate, could feel too scripted. “It would guide people to the right place, but it wasn’t very conversational,” Cook says.

The team introduced agentic AI to better understand customers’ natural language and respond more conversationally. Working from a single Talkdesk Retail Experience Cloud instance, the team built its AI agents in-house, using a supervising agent to coordinate four specialists focused on returns, warranties, frequently asked questions, and product questions. Three customer-facing agents support the Tuft & Needle, Serta, and Beautyrest storefronts.

As the team applied lessons from each launch, implementation became faster, shrinking from six or seven weeks for initial launches to just a few hours for the Beautyrest agent.

Knowing when AI should step back.

Not every customer interaction should be fully automated. Mattress returns can involve coordinating pickup partners, delivery dates, and different processes depending on the product.

Serta Simmons Bedding used its experience and thousands of customer conversations to establish clear rules for when AI should gather information, guide customers, or hand off to a human agent.

For returns and price adjustments, AI collects relevant details, including product type, purchase channel, and reason for contact, before directing customers to the appropriate next step. For price adjustments, it captures the order number and passes the request to an employee authorized to issue the credit.

Customers are informed when they’re interacting with AI and can request a live agent. Refunds and price adjustments remain under human control.

This approach keeps routine interactions moving while ensuring customers can reach a person when their situation requires additional support.

Measuring success by how customers feel.

The impact was visible within the first week of the Tuft & Needle launch. Chat customer satisfaction increased by more than a full point, and Serta showed a similar trend after launch. The AI agents also achieved 96–100% confidence in identifying common customer intents.

Serta Simmons Bedding doesn’t measure success by automation alone. Customer satisfaction is a launch gate, helping ensure the experience is genuinely useful rather than simply moving interactions through a system.

During the first two weeks after each brand launches, the team reviews every AI-handled conversation, identifying gaps and refining responses. When something goes wrong, team members follow up with the customer and work to make it right.

“‘Human in the loop’ gets thrown around a lot. For us it’s simpler—when something goes wrong, we own it and we make it right. That’s the humanity we bring to customer experience,” Cook explains.

Giving agents more time to make a difference.

Serta Simmons Bedding is extending AI’s impact beyond self-service with Talkdesk Copilot. Deployed across its business-to-business (B2B) and direct-to-consumer (DTC) teams, Copilot’s call summarization capabilities have reduced after-call work by more than 50%.

Copilot creates structured summaries capturing the reason for contact, key details, actions taken, the outcome or next step, and customer sentiment. Agents spend less time documenting conversations and more time helping customers.

The company is also using Talkdesk Navigator to improve call routing. On DTC lines, refinements to caller prompts reduced the no-match rate from approximately 50% to the high 20s and low 30s. On the dealer side, a new IVR flow reduced misdirected calls by 35% within two days.

Attribute-based routing further connects customers with agents whose skills match their needs, whether they’re asking pre-purchase questions or seeking warranty support.

Continuously improving the customer experience.

Serta Simmons Bedding continues to expand its use of Talkdesk across quality management, routing, and workforce operations. The team is introducing generative AI evaluation through Talkdesk Quality Management and plans to extend attribute-based routing to dealer support.

By building and refining its AI agents in-house, the company can respond quickly to customer feedback and apply lessons across its brands.

For Serta Simmons Bedding, AI isn’t about replacing the human side of customer service. It’s about resolving routine questions faster, giving agents more time to handle complex situations, and making every interaction more helpful.

The result is a more responsive customer experience that brings together the speed of automation and the empathy of human support.