How CUURT reached 50% resolution with Engaige in two weeks

CUURT is a direct-to-consumer brand selling court shoes for tennis, pickleball and padel, mainly in the United States. The company is one year old. The core team is six people, the creative team eight including two freelancers, and support runs on Shopify, Richpanel and Loop Returns. Customers reach support through email, chat, Instagram and Facebook.

The product is built around one idea: a wider toe box that lets the toes spread naturally. Players can stay on court longer without fatigue or pain, and that is one of the main reasons they buy. CUURT owns the design, works closely with its manufacturer and ships improvements roughly every quarter.

The brand exists because the founders could not find what they needed themselves.

We were looking for a solution that didn’t exist for the sport we were playing, tennis in particular. So we had to make it ourselves.

The US was a deliberate choice. One language, one website, one set of ads, and a market that rewards you if you can crack it. The next steps are other English-speaking markets such as Australia and the UK, followed by a broader test in Europe.

400 to 500 tickets a day, and a team that could not catch up

Before Engaige, CUURT’s support was almost entirely human. Seven to eight agents worked in Richpanel, handling 400 to 500 tickets a day. Returns and exchanges were the biggest category, followed by “where is my order?” and order changes. The team knew other brands were already using AI for support.

We knew other people were implementing AI, and we were a bit behind.

Two attempts before Engaige

The first step was the AI built into Richpanel. CUURT tested it for a month and was not convinced. The founder wanted something more customisable, and something independent of the help desk itself. So he started building his own. A set of AI agents in Claude, scheduled to run every few hours. It worked, up to a point.

I thought it was easier. But it got messy very quickly, because it’s difficult to really see what the agents are doing and what their reasoning is.

The hard part was structure. Which actions exist, which policy applies, and under which conditions a human takes over. Every time an agent failed, a hard rule sent the ticket back to the team, so the humans still did most of the work.

With Claude you ask for this edit or that edit and it implements it, but you don’t know where the change is actually going. With Engaige you know it’s in this policy, in this step. It’s cleaner, and it’s less risky.

Met at an event, demo the same day, live the week after

The switch came through word of mouth. At an event, the founder met someone who was using Engaige. He booked a demo with Sander, Engaige’s Chief Commercial Officer, the same day and kicked off on the platform the week after, on 25 August.

Sander and Aaron, Customer Success at Engaige, ran the onboarding. It took at most two calls in the first week. After that, CUURT was already setting up its knowledge base and policies, testing them and switching them on for live tickets. A shared Slack channel with Sander and Aaron covered verification and the custom actions CUURT needed. “It was pretty cool to see the first tickets being automated.”

Today the team runs the platform almost independently. The founder is candid that optimising it takes work, but that was never the problem he was solving.

It’s easy to understand. It’s easy to see what is happening and why it is happening. That is what was lacking in the system I built myself.

Richpanel stays, Loop Returns and the 3PL plug in

CUURT kept Richpanel as its help desk. Engaige works on top of it, connected to the systems that hold the answers. The Loop Returns integration lets the agent check whether a return exists and what its status is. When needed, it sends the customer a return link that opens their order already logged in, so they can make changes themselves.

For shipping, CUURT built a custom integration with its 3PL, Day1, together with Aaron. Day1’s API shows a shipment’s status before it reaches Shopify, so the agent can tell a customer with more accuracy whether an order has shipped, is out of stock or has a problem.

An emergency approach: reply first, perfect later

Most brands start with one ticket type and build up. CUURT did the opposite. All policies went live at once, with the risky actions, refunds and order edits, still handed to the team.

The reason was the backlog.

The support team couldn’t catch up with the number of requests. So we preferred to aim for a less accurate reply, which can still be good, but we want to reply to the customer anyway. Then we have time to optimise and improve. It was an emergency approach.

Since then the boundaries have moved. Engaige now handles almost every action on its own: order edits, partial refunds, and cancellations within the six-hour window before an order is fulfilled. Full refunds are the one action that still goes to a person.

Fewer tickets in, cleaner tickets for the AI

Alongside the AI, CUURT rebuilt its help centre so customers can solve more themselves. The founder sees that as the step that comes before automation, not instead of it.

The most important thing, even before AI, is to let the customer solve the issue themselves. They have a better experience. They don’t have to contact support, reply to different emails or even type.

The effect showed up inside Engaige too. Fewer tickets, less noise, and cleaner tickets that make the agent easier to optimise.

The results so far

Engaige resolved 40% of CUURT’s tickets on its own straight after set-up. Two weeks later that was 50%, and the team is now working together with Engaige towards 60%. For the tickets that still go to a person, the team’s first reply time went from 19 hours to 3.5 hours in the first two days. The agents are under less pressure. They now spend their time on the cases that need judgement.

CUURT is also honest about what the numbers do not show yet. The backlog of recent months pushed CSAT down, and that recovery is almost complete. The next target is a human reply within two hours, and with it, higher satisfaction.

Humans stay in the loop

Automation does not mean the team disappears. With Engaige taking on the volume, the agents focus on the tickets that need a person, and the founder expects it to stay that way.

Even if we reach a higher automation, there will still be tickets that are tricky, new or unusual. We still need humans.

What comes next

The focus now shifts from recovery to optimisation, and the volume keeps growing: CUURT expects 200,000 support tickets a year. The next targets are 60% of tickets resolved by AI and a human reply within two hours, so satisfaction can climb back up.

On the commercial side, Q4 is mainly a marketing push: stock was planned months ago, and the creative team is growing, especially with video editors.

Further out, CUURT wants to reach more demographics in the US before expanding to Australia, the UK and a broader test in Europe.

What CUURT’s founder tells other founders

Asked what he would say to another e-commerce founder in his old position, CUURT’s founder does not hesitate.

Don’t build the agent yourself if you’re aiming to automate most of your tickets. It takes a lot of time and it won’t make sense, financially or time-wise. It’s better to use a platform like Engaige that has already built the infrastructure. It’s more reliable, and it’s interpretable. That’s very important.

Ready to find out how much of your support already answers itself?

If your team is falling behind on tickets, book a demo and we will show you what Engaige resolves on your own tickets. You keep your help desk, whether that is Richpanel, Gorgias, Zendesk or something else.