Intelligent Dialog Systems in Client Management: Faster, Better, Smarter
What will the client communication of tomorrow look like? Daniel Welzer, CEO Arvato CRM Germany, discusses the trend toward intelligent dialog systems that learn autonomously on all channels.
Daniel Welzer, CEO Arvato CRM Solutions Germany
Many hotlines today use a semi-automated speech dialog system called an IVR system in client dialog. Is this the future of client communication?
Welzer: I hope not. Most clients find it annoying repeatedly entering numbers before finally getting to speak to an advisor. It’s not practical – for the client or the company. That is why we at Arvato CRM Solutions focus on complete automation, and why we have started a partnership with IBM Watson.
What are you working on with IBM Watson?
Welzer: We’re currently taking a closer look at past client communications as part of a pilot project. Specifically, we are feeding these communications into a cognitive system via an interface so that it learns about the structure and topics covered in our dialog with clients. Beyond that, there are analytical methods for understanding language as well as other methods like neural networks involved, which are crucial to making the right decision based on past information. This will enable us to automate dialog in future. This could be done with a chatbot, for example. Using analytics, a human will only come into play if the system determines the conversation is becoming too complicated. This will also lead to greater acceptance by customers. They will be able to describe their problem to the chatbot without wasting any time when their request is being presorted.
If the technology is already so advanced, then why are there still so many hotlines robbing clients of their time and patience?
Welzer: Though we haven’t yet reached the point where we can roll out this technology comprehensively, we’re not far off the turning point where the whole development will start gathering pace and different technologies will start to interact. The quick pace at which technology is now progressing and the growing capabilities of analytical systems are what will make this possible. Furthermore, the price of technology will decrease as its use increases. The technology will also help to rectify the skills shortage in client management since automation means that ten to 20 percent of customer service queries can simply be digitized. Employees can then use this additional time to concentrate on the important job of advising clients.
Nowadays clients are active on many channels: They phone, email, text and post. How do you navigate this?
Welzer: There is a trend toward omnichannel communication. The trick is to merge each channel so seamlessly that no knowledge gets lost, or even to use different channels in parallel. The new platforms that we are now introducing orchestrate these channels simultaneously and do so significantly better than their predecessors. This is the icing on the cake for large-volume businesses.
Cloud services are still very rarely used in client management. Why?
Welzer: I have noticed that cloud offers are growing, but at a fairly low level. But if we want to network channels, monitor the mobility of clients and play internationally, a cloud solution is vital. The question is simple: How secure do I make my cloud? Since Arvato CRM Solutions operates its cloud service in Germany, the data doesn’t leave the country, making it safer. Clients appreciate that. Yet there is still a large backlog in demand for cloud services overall.
How do you see the pace of development of cognitive technologies?
Welzer: In the past few years there has been an enormous leap in development, allowing for entirely new solutions. The next major step would be if data, e.g., dialog data, could be automatically and continuously read out with the associated results. One simple example of this would be a chat dialog between a client and a service employee that could be linked directly to how satisfied the client was. With this type of linkable data, a machine’s learning process becomes virtually automatic, resulting in maximum speed. Depending on the system environment, most categorization must be done manually, which restricts the machine’s speed to that of the person operating it.
What do you want from these technologies? What is still lacking?
Welzer: It would be great if we could drive this cognitive technology forward faster and thereby accelerate the leap from the pilot project to a wider platform. But this depends on client devices – from browsers to smartphones and tablets. Future platforms should work on all devices. That will require a higher degree of standardization. The next ten years will certainly be exciting, especially once machines are capable of learning automatically and becoming more intelligent. How will humans handle this? Who will decide what is right and what is wrong? The human or the machine? Although a machine’s decision could at first seem unusual to a person, it may in fact be a better decision than a human’s.
Author: Editorial team Future. Customer.
Image Source: istock.com/NicoElNino