welove.ai header image

Can we add some empathy where it’s needed most?

A sales offer can be a pleasing thing if it fits and comes at the right time. But it's not always the right time to make an offer, is it? The wrong offer or even the right offer at the wrong time ruins much more than it can ever achieve. Especially when it comes to a text chat dialogue. As such, we believe that empathy also means knowing what is better left undone.

{fCan we add some empathy where it’s needed most?

Chats have come a long way – from the first SMS to highly evolved chat tools like Apple messages or the ultra popular WhatsApp. It looks like these chat clients hit the right spot in fulfilling mankind’s desire to communicate. More than 100 billion messages a day are sent – and that’s just on WhatsApp. The younger generations in particular jumped on these channels and quit using their phones just as telephones, turning them into texting engines. Given the importance of this real-time text communication, it is obvious that we have to use it in our communication with customers, too. However, these real-time channels also pose a challenge. Because of the lack of facial expressions, gestures, and tone of voice, such communication can tend to be complicated and generate misunderstandings. In text conversations, it is difficult to express the right emotions. This is why emojis were invented, but they only help up to a point. Quite often you have to just guess what the sender really wants to say. And that is only the human side of this kind of communication – things get even more complex with chatbots. Nevertheless, both WhatsApp and chatbots are key components of our customer communication. And not to forget: Every chat contact is a potential sales opportunity. So how can we get it right?

How we teach empathy to a bot.

We start small and safe at first. We know that people use Frag Magenta on telekom.de or our WhatsApp channel mainly for support purposes. That means they are obviously not around for a sales offer in the first place. It doesn’t take a lot of empathy to understand that – common sense is enough. So what we do is that we teach the chatbot to come up with the most useful solution for the customer’s problem. The bot has machine-learned from our data to identify what the customer’s issue is, so it can come up with proper solutions. And if the bot has done its job well and the customer’s mood is then presumably good, a suggestion for improvement can be made, right? But how do we know if the time is right? The procedure follows a plan: Did we solve the problem? Yes. Is the customer satisfied? Ask them – if they give their approval with a “five star rating”, we suggest what might be interesting for the customer, as derived from the preceding dialogue: a faster DSL line, a new router, or maybe the latest iPhone. Under the hood of all this is a so-called NLU engine, which means “natural language understanding”, and our “One Conversation eMpathic AI”. The latter generates “next best actions” – those actions, in this case offers, that best match the customer’s current situation and needs. Could you call this digitally empathic? We definitely think so! Stay tuned for our next inside story on this channel.


One Conversation. A program for an empathic and reliable customer relationship.

One Conversation is Telekom’s new program that will guarantee a consistent and positive customer experience across all thinkable touchpoints. To this end, an AI-based mind is being built up to ensure that our dialogue with the customer is a continuous conversation. No matter whether we chat with them, talk to them on the phone, or meet them in the store. We will be as familiar with our customer as with a long-time friend and know what they like and even what they don’t like. This is how we cultivate an empathic relationship with our customers and how we always know what the next best action will be.

Stay tuned.

Yours, Jan