Chatbots and artificial intelligence: What is already possible today?

by | 19 October 2020 | Basics

Chatbots owe their success and relevance primarily to artificial intelligence. Speech, gestures and facial expressions are the simplest and most natural forms of interpersonal communication. So what could be more natural than using natural language to communicate with machines? The first attempts to use natural language as a communication tool to control machines are therefore not entirely new. At the beginning of the development, however, speech recognition and voice control did not really make customers' lives easier. The following commercial from 2008 takes aim at precisely this aspect:

Even if they weren't called that at the time, the earliest chatbots were based on a simple system. At best, they could identify individual keywords and perform a specific action linked to them. Through the developments in the field of Artificial intelligence and in NLP (Natural Language Processing), chatbots are now much more capable of being an intelligent interface between people, machines, information or services.

What is a chatbot?

Even though there are several names for chatbots - digital assistants, virtual assistants, chatbots or AI assistants - so it can generally be said that it is intelligent software that focuses on talk to people in a natural way can.

Chatbots and artificial intelligence can be integrated into existing infrastructures: Whether in a Connected Caran app, a homepage or a messenger service such as WhatsApp, Facebook Messenger or Slack. There, chatbots form an interface between people and specific information or services.

Are chatbots already accepted in society?

A representative Survey by the industry association Bitkom found that a quarter of all Germans can already imagine using chatbots today. Why? Above all, the short waiting time for answers and the support with scheduling, research make the use attractive.

2018 saw the arrival of a Dentsu Aegis study concluded that chatbots have not yet reached the centre of society. The main problem: much Too few people have come into contact with chatbots so far. According to a Analysis by the management consultancy Lünendonk the lack of skilled workers is one of the main reasons why companies do not have more digital solutions like chatbots in place. According to this, it is not the underlying technology and experience that is failing.

According to the latest Study 2020 by Unymira the trend has not really taken hold in the area of customer service. According to the study, only 1/3 of the respondents have a chatbot in use. It is exciting that almost 70 % are planning a bot project in the future and 26 % of them even in a short time window.

Two basic manifestations

As already indicated, there are very different types of chatbots. Similar to communication with humans, we can chat with bots via Messenger or simply converse with them in the form of virtual assistants such as Alexa or Siri.

Narrow field of activity

At the moment, two general development trends can be identified. On the one hand Very specialised chatbots developed, which is based on a Narrow range of tasks or even tailored to a single task. For example, there are:

  • Weather experts
  • Assistant for medical diagnoses
  • Helper to order an Uber taxi
  • Deliverers of current news according to personal preferences
  • Specialised marketing chatbot in the business environment

Wider range of functions

On the other hand, however, there are technologies that attempt to Wide range of different functions cover. Assistants such as Siri, Cortana, Alexa and Co. are available to the user as omnipresent - and above all omniscient - digital helpers: From the weather forecast to appointment and travel planning and shopping assistance to the control of a smart home, chatbots should be able to serve all the needs of the user.

Even if the degree of maturity of these assistants currently leaves a lot to be desired, the positive trend is clearly recognisable. In particular, the large technology companies Amazon, Apple, Google, Facebook and Samsung have been developing assistants with this basically unrestricted scope of purpose for some time now. Strategically important business area recognised.

Areas of application of chatbots

The areas of application for chatbots are diverse. Nevertheless, they are better suited for some areas of application than for others. Especially very conversation-intensive and emotional topics are better handed over by bots to colleagues in the service centre. According to the current unymira study, chatbots are most frequently used in the areas of:

  • Standard information
  • Surveys
  • Support
  • Pre-qualification
  • Order processing
  • Consulting
  • Complaint

for use.

Chatbots areas of application
© unymira

Chatbots vs. Artificial Intelligence

Due to these different manifestations, there is always a lack of terminological clarity. The terms are often mentioned in the same breath and in fact there are differences between the two. Both technological developments have numerous links. Nevertheless she not synonyms. Chatbots are the end product or application, while artificial intelligence is an underlying technology that works in the background.

The chatbot "Karl Klammer", an animated paper clip that provided access to the help pages in Microsoft Office until 2004, still managed completely without artificial intelligence. Depending on which keyword appeared in a question, it searched for suitable answers from the linked help file.

Even today, not all bots are necessarily based on artificial intelligence. The Slack bot familiar to every Slack user functioned without AI until recently. Simple applications, questions or tasks can be covered by linking commands, commands and keywords with the associated databases. In the future, however, developers will hardly be able to do without the advantages of artificial intelligence in speech recognition and processing.
If you want to know more about the trend and the risks and opportunities of this technology, read more in the Interview with Marius Försch.

Golf, Golf or Golf? On difficulties of understanding and the ambivalences of human language

One of the biggest challenges is to find the Ambiguity of human languagee. Sometimes we use ambiguity intentionally, for example when flirting. But often the semantics are unintentionally ambiguous. For example, the word "golf" has several, very different meanings. It can be a sport, a car or a geographical formation (like the Gulf of Mexico). We can derive what exactly is meant in each case relatively easily from the context. For chatbots without artificial intelligence, however, it is difficult to understand such contextual relationships to understand.

However, natural language also has many other ambivalences and sources of misunderstanding. Colloquialisms, dialects or irony can make it difficult to understand. Even simple spelling mistakes can make it difficult to understand what is meant.

Chatbots that give too brittle answers are not as accepted by users as chatbots with a touch of humour and naturalness.

Regardless of whether the natural language is in spoken or written form, it can be Artificial intelligence help to eliminate difficulties of understanding such as these. In particular through Deep Learning, a technique in the field of artificial intelligence, Chatbots are becoming increasingly intelligent.

It is important to know that chatbots become more intelligent the longer they are in use. They learn through the User feedback constantly and improve their skills as the time of use progresses. Statements that may not be understood the first or second time around are soon no longer a problem thanks to user feedback.

Bots and the end of the app era

Because of skills like this, today there is always talk of the End of the app era the talk. Chatbots and artificial intelligence are the next "natural" step in a multi-step evolution. First, every company had to develop a Homepage where customers and interested parties could get an overview of products or contact the service.

In the second stage of development, it was good form for manufacturers and providers to offer one or more apps. The ambitious goal of companies like Apple, Google & Co: To replace the Apps should in future Chatbots step. Instead of clicking through menus and subpages on homepages and filling out forms, you simply tell the chatbot what you want to know or order. Convenience and Simplicity are the main reasons why chatbots are predicted to have a great future.

Chatbots: Opportunity for companies and customers

The question of whether chatbots will actually be a success or just remain a short-term hype is no longer an issue. Chatbots have great potential and are being implemented more and more frequently in companies. They are becoming increasingly important, especially in customer service.

Nevertheless, developers must be able to successfully integrate chatbots into their existing infrastructure. However, customers do not just expect chatbots that simply work. For them to be accepted, it is necessary that they are practical, smart and user-friendly are. At best, they also have a little Humour.

If all this succeeds, chatbots bring with them a whole range of benefits. Chatbots do not sleep. Customers can reach a chatbot around the clock. Many simple questions can be answered in this time. Moreover, chatbots know no capacity limits. Even if 100 people enquire at the same time, every customer is served immediately without waiting. Annoying waiting with bad music is thus a thing of the past. If it is more difficult than expected, the chatbot automatically forwards the request to a member of staff.

In the best case scenario, a chatbot knows its counterpart, which enables a Personalisation of the Customer Journey is made possible. When I order my pizza via chatbot for the 5th time, it starts to learn my personal preferences. If I forget, the chatbot might remind me that I usually order with extra cheese.

The inhibition threshold to call a hotline with a seemingly "stupid" or naive question, where a caller may have to wait 10 minutes, is also relatively high. A Chatbot does not judge about the questioner, so that even pure comprehension questions are likely to be asked much more frequently. The consequence: customer satisfaction and trust increase.

In a recent Study by unymira 180 experts and managers in the DACH region were surveyed on the topic of "Chatsbots - the path to transformation". We have summarised the three key findings for you.

Chatbots facts
© unymira

The most important advantage of chatbots and artificial intelligence results from the way they are made. Chatbots are simple to operate. They understand and master human language and thus give the user the feeling of a natural conversation - without interaction interfaces such as forms or homepages. For this reason alone, chatbots and artificial intelligence represent a great opportunity - for companies and customers alike.

<a href="https://www.alexanderthamm.com/en/blog/author/marius/" target="_self">MARIUS FÖRSCH</a>

MARIUS FÖRSCH

Marius studied at the Technical University of Kaiserslautern and managed data & AI projects at Alexander Thamm GmbH from 2016 to 2018. He is currently a freelance Data Product Manager & Strategist.

0 Comments

Submit a Comment

Your email address will not be published.