Direct-to-consumer mattress brand Saatva streamlines customer support
Saatva is ensuring its customer support program adequately resolves issues as the company expands and shopper needs evolve.
Premium direct-to-consumer mattress brand Saatva has grown significantly since launching as a digitally native, online-only retailer in 2010. This includes expanding to the brick-and-mortar channel in 2019 with an initial physical showroom in Manhattan.
Now operating 33 brick-and-mortar locations with seven more scheduled to open by the end of 2026, Saatva has discovered over the years that its customer call center required upgrading and more sophisticated technological capabilities to meet the evolving needs of its customers and its business.
Marty Ehrlich, chief customer service officer of Saatva, recently sat down with Chain Store Age to explain how the DTC brand has leveraged artificial intelligence-equipped technology from Zendesk to help ensure customer support can adequately resolve the needs of its shoppers.
What customer support issues was Saatva experiencing?
For Saatva, it wasn't a matter of issues, it was a matter of growth. We started with Zendesk around 2016. Previously, we had been able to handle customer support through emails and spreadsheets, but as our business evolved and grew, we knew this model wouldn’t give our customers the level of service that they needed.
We also wanted to make sure that we had a system that would show customer support personnel exactly what was going on so that they could see notes and the company could communicate internally using a ticket that contained all the history with that customer.
Within a few years after starting with Zendesk to assist email customer support inquiries, we were able to link in our phone system, where phone calls created tickets, also providing more visibility into each customer's journey. This allowed us to know which representatives assisted the customer.
Saatva customer service is open 24 hours a day, seven days a week. Customers can get a live person or AI chat.
[READ MORE: Survey: Incomplete customer support experiences can hurt brand credibility]
What Zendesk customer support solutions are Saatva using?
We are using Zendesk AI auto routing, which has reduced our first touch time by more than 60% so we don't have to have a team of people reading each incoming ticket and deciding who would be the best person to handle it. AI does that.
Saatva also uses the Zendesk AI chat tool. One of our hallmarks is that we want to communicate with the customer how they want and when they want. With live chat, we let the customer know immediately that they are interacting with a chat bot and that they can also chat with a live human being. About 90% of our customers decide that they want to start the customer support journey with the bot.
There is also a quality assurance solution with some AI functionality that helps ensure our representatives are providing the customer support experience that we expect them to give. Previously, we had a team that every month manually inspected about 10 to 15% of the tickets just to make sure that if macros were selected, they were appropriate; or that written responses used good grammar and were professional with the right tone.
The Zendesk quality assurance tool now gives us the ability to inspect 100% of tickets, which has been terrific. In addition, we use an agentic AI product that will recommend a macro based on the specific customer support situation. The agentic solution can also suggest previous tickets with a similar situation, which helps get newer reps up to speed faster.
Our goal is never to have AI replace the human interaction. We want to augment the customer service rep to enable a better customer and employee experience.
What has response to the customer support technology been, from both customers and employees?
With the exception of AI chat, the technology is completely behind the scenes so customers are not aware of it. However, our most recent monthly customer satisfaction score was 4.92, so we are excited about that.
As I said previously, 90% of customers who enter the support protocol choose to chat with the bot as opposed to a human, and about 67% of the time the bot is successfully able to resolve the situation.
Are there any future plans for your customer support center you can discuss?
It would be great for every rep to have an AI assistant that would let them get information out of our knowledge bases or procedures a little faster. We may lightly test some voice bots overnight and seeing what we learn from that. Saatva is open to learning about and testing anything that might enable us to provide a better customer support experience.
