Do you enjoy waiting on the phone for 40 minutes?

Between humans and robots, I prefer humans. Most of the time, at least. But when I’m on the phone for 40 minutes waiting for my call to get picked up, I just want to get the whole thing over with.

Here’s a twist to that story: what if the voice sounds human, but isn’t? What if it’s generated by AI, navigating through my queries in real-time and successfully solving my problem? Is that enough to throw all my “preferences” out the window?

I’d rather not enter the moral debate of machines over humans, jobs being replaced by AI and so on; that’s too complex to address here. But I will say this: preferences be damned, so long as I can cancel my pointless pet insurance! While I prefer to hear a human voice on the other side of the line, ultimately, so long as my query is resolved, I don’t really care.

I have a suspicion that I’m not alone. Talking to Talkr.ai at CES, I get the feeling the French company figured that out almost a decade ago! Based in Levallois-Perret, right in the heart of Paris, the company has made it its mission to develop automated customer service assistants that are as modern as it gets. And it has wasted no time leveraging AI and machine learning to make them even better.

Interview below, edited for clarity.

Could you please introduce yourself and the company?

I’m Katya Lainé, CEO and Co-Founder of Talkr.ai.

Our two main offerings are a no-code platform that allows businesses to create their own virtual assistants with ease and we also offer Assistants-as-a-Service, which are plug-and-play, off-the-shelf, multidomain assistants that help companies interface with different customers or audiences.

An example of our products in action can be seen in contact centres, where we deploy virtual assistants that handle queries over the phone. That happens to be our main product, driven by the fact that 50% of interactions with the brands happen over the phone. While our solution offers omnichannel capabilities, voice is very important to us.

Over the phone, our assistants pick up 100% of all conversations, 24/7, with no wait, and resolve between 35% and 85% of queries That’s a very powerful layer of automation without any human interaction.

35% to 85% is quite a high variance – what affects it?

It depends on the use case. For example, one of our clients in transportation and the query they often receive is “Where’s my delivery?” For that use case, we can automate and resolve 85% of queries. In other cases, such as in healthcare, we have virtual systems that aid with patient management. Even in those cases, resolving 35% of queries is still a pretty big feat.

When a call comes in, is it an AI-generated voice answering back?

Yes, it’s done through AI. The system we’ve built orchestrates voice recognition, natural language processing and then, of course, AI-generated voice. This is what enables us to address the query in real-time and immediately provide an answer over the phone.

How’s business going for Talkr.ai? How long have you been around?

We are a pioneer in France and possibly the world. The bot we’ve developed and the platform it’s part of have been in the market for around eight years, well before ChatGPT and all the excitement that followed. Our first automated conversation over the phone took place in 2016.

Now the market has caught on and we’re very excited because we are in a position to capitalise on the increased demand for automation.

What’s your customer base like?

Most of our customers are in Europe, but we have many in the US. Our solution is multilingual and is currently available in Dutch, Bulgarian, Spanish, English and a few others.