I thought 50% was way too ambitious, but now I don’t even think about our chat channel anymore

Together with our very first pilot customer, Bikeshop, we handled 56% of all incoming chat inquiries in August. And when we say handled, we mean handled.

That means the customer received a correct answer, with no need for a human to follow up later.

We’re not just talking about simple questions like “What are your opening hours?”, but real assistance. Bikeshop gets questions ranging from how to cancel or track an order, to which axle they need to use a bike trainer they’ve been considering.

High Volume on Customer Support

“We’re not a big team in customer service at Bikeshop, and before, we typically needed 1–2 employees fully dedicated to chat, while the rest managed phone and email. That left us with very little capacity,” says Niklas Johansen from Bikeshop.

Bikeshop is committed to great customer service, and providing good answers requires attention to each customer. On the busiest seasonal days, that used to mean long wait times.

“Typically, there was at least a 30-minute queue on the chat, but now I hardly even think about the chat anymore.”

Didn’t Believe It Would Work

When CognitionHub first approached Bikeshop, their IT manager, Dan Morgan Karlsen, wasn’t particularly convinced it would work.

“50% sounded very ambitious, and we get a lot of difficult, technical product questions. There are so many garbage solutions out there.”

Still, they thought the price was more than worth it if it actually worked—so they took the chance. And it paid off.

“The bot has brought us significant value—easily equivalent to at least a 50% position as it stands today.”

Now, customer service at Bikeshop can assign one person to chat, who can also handle calls and emails, because the chat volume is almost gone. The rest of the team can share other inquiries and focus on email.

“This has saved us a lot of time, and now the customer service team even has the capacity to help out in the store when needed, which is really valuable.”

Happy Customers

The feedback Bikeshop has received has been positive, and customers have purchased full bike kits based on the bot’s recommendations. Product recommendations have been a major focus to make the bot outperform anything currently on the market—and that’s shown real improvement.

“Customers are used to bad chatbots—there’s a lot of junk out there. But this one actually works.”

The Road Ahead

Going forward, we at CognitionHub will continue improving the product and user experience. One of our upcoming priorities is to also reduce the pressure on email support.

Talk soon!