CallDesk raises $2.5 million for its AI agent for customer support calls

Next Story

HMD’s revived Nokia 3310 classic mobile gets 3G

It’s time to put an end to these dumb automatic customer support calls that ask you to press 1, then press 4, then press 2, then say “yes”… CallDesk is working on a virtual agent for customer support calls. Think something like Siri for customer calls. Point Nine Capital and EQT Ventures are investing in today’s funding round of $2.5 million (€2.1 million).

There are multiple components that make CallDesk stand out. First, the company has been working on natural language processing capabilities so that you can have a normal conversation with your virtual agent. It starts by understanding what you mean even if there are dozens of ways to say the same thing.

Second, CallDesk understands context. For instance, if you’re trying to book a dentist appointment, CallDesk knows your answer to the previous question. This way, you can say something like “what about a bit later on that day?” and CallDesk knows what day you’re talking about based on your previous answer.

“Our average total discussion success rate is 80 percent and can go up to 95 percent for simpler use cases,” co-founder and CEO Vincent Gire told me.

The company is starting with low-hanging fruits. The virtual agent can handle authentication for banks, order status for an e-commerce platform, call routing, scheduling and more. The idea is that CallDesk won’t replace all your employees in your call center. But maybe they can just focus on the most difficult tasks. It can also help when it comes to reducing waiting time.

CallDesk is a software-as-a-service platform and relies on an API to route calls and integrate with existing call center solutions. It integrates with CRMs or ticketing systems.

So far, the startup has attracted 10 big enterprise clients that pay a price per minute. Customers include La Poste and MGEN. CallDesk already supports 54 languages.

With today’s funding round, the company plans to give more control to its customers to that they can tweak their virtual agents to handle more tasks.