Digital assistant in practice: Successful implementation in companies

from | 15 November 2018 | Basics

What can a digital assistant already do today? Anyone who has ever had a conversation with Siri or Alexa knows that one moment of frustration. As soon as Siri doesn't understand something, at best it only shows Google search results. Even though digital assistants are getting better and better, there are still limits to what is possible. This is especially true when users expect their counterpart to be intelligent in a similar way to themselves. This idea is problematic, however, and makes for the aforementioned frustrating moments.


By loading the video you accept YouTube's privacy policy.
Learn more

load Video

The status quo in M2M communication: Even if the attempt, Chatbots having a conversation with each other does not yet lead to a particularly deep or consistent conversation - some responses are surprisingly humorous and appropriate.

Digital assistants are not yet all-rounders. There is still a lot of development work to be done before we will be able to converse with them naturally. But that, too, is part of taking stock of the status quo of digital assistants: The functioning Use cases make life much easier for users. Anyone who simply wants to know what the weather will be like, how many inhabitants there are in city X, add something to their shopping list or send a message will be able to complete these tasks much more quickly with the help of digital assistants such as Siri, Alexa & Co.

Users will experience all this as a sense of achievement if they do not expect more from their digital assistants. For companies that use digital assistants, this means that they must communicate exactly which services they actually offer with them. Since digital assistants do not fulfil a pure end in themselves, companies need to establish a relevant added value and also be able to comply with them.

Practical areas of application of digital assistants

In order to be able to communicate to customers and interested parties exactly what the purpose of digital assistants is, it is first necessary to clarify what is possible and useful in practice. In the Marketing for example, they can play an important role. Prospective customers looking for information about a product or service do not have to laboriously click through search results, but simply ask a bot what they want to know. This not only makes the search easier for customers.

Rather, companies learn much more precisely over time what knowledge they need to provide. Since digital assistants have the ability to recognise questioners at a later point in time, they can be much more precise. better meet their needs and communication can be personalised take place. If the interest condenses into a concrete purchase request, the bot can forward the customer directly to the shop.

Digital assistants can help customers along the entire Customer Journey accompany. In the sales area, order processes can be personalised, making the sales process more customer-friendly. Because the bot knows the questions a customer has asked beforehand, it can identify the customer's individual needs and special features and thus create a personalised proposal create.

Additional features, other suitable offers from the range or recommendations on insurance can be customised. Even after the purchase, they offer the customer many advantages. Customers with questions about their products don't have to go through all the FAQs or browse through forums. They ask the digital assistant their specific question and if no suitable answer is available, a human service employee is notified directly. This can significantly improve the service.

The central advantages

The Personalisation of the customer journey is one of the most important features. Hardly any other method offers such a high degree of individualisation. Since they are integrated in messenger services or via other communication platforms, they know their counterpart. They can address customers by name, recognise them, know their particular needs and preferences. Digital assistants can incorporate all this information when advising and responding to requests.

Another central advantage of digital assistants is the Naturalness and Simplicity of the exchange they enable customers to have. Thanks to artificial intelligence, they can be programmed to understand natural language (written or spoken), they are enormously easy to use. This improves the customer experience and expands the customer or user base.

Employees don't have to ask their secretary for the current Wi-Fi password every time, but simply ask the digital assistant. And anyone who uses a chatbot to find out how much a taxi ride to the airport costs can order a taxi in the same breath.

The digital assistant is an open systemor a "single point of truth". No matter what I need, I can turn to the bot. They are not bound to opening hours, but are always available. They don't have any closing time and can therefore expand the range of services offered by companies because they are available around the clock. The scope of what is possible with them can be determined by each company itself. However, this is not enough to create a successful chatbot that customers like to use.

The right mixture makes the difference:

Digital assistants can fulfil very different functions. However, in order for them to be used successfully in companies, the focus must first be on three fields. In the run-up, it must be clarified both which User experience be offered, which Knowledge and what information and with which Technology the digital assistant is to be implemented. Digital assistants must be viewed from these three different perspectives in order to make them an effective tool.

In order for the digital assistant to be accepted and used by customers, they must first be considered from the point of view of the User experience be considered. This includes determining the individual user stories. The individual use cases are then derived from this. The development of the bot's personality is a factor that should not be underestimated. A central, success-critical factor is that the assistant has a sense of humour and gives the impression that it is a human counterpart.

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

However, the most sparkling answers are of no use if the information provided is irrelevant or incomplete. That is why the second step must be to Chatbot with all relevant resources or sources of information to link. To do this, the question categories must first be determined to then ensure that the best possible content can be provided. However, the mere identification of content sources is not enough. All content must also be "botified" so that a chatbot can output it in a meaningful way.

Last but not least, a Chatbot also as technology be considered. How intelligent should or must a digital assistant be in order to cope with the tasks determined in advance? The products do not necessarily require the development of complex neural networks, even if this makes sense in other cases.

In many cases it is enough to Open Source Tools which are also based on neural networks under the bonnet, but do not require any knowledge in this regard for development. In addition to choosing the right software, there are other technical requirements, especially for large companies. For example, they must be integrated into the be integrated into the existing IT architecture. Legal requirements must also be checked and complied with. Not everything that is feasible is compatible with data protection.

A chatbot can only give certain answers if it is equipped with sufficient intelligence.
A chatbot can only give certain answers if it is equipped with sufficient intelligence.

How intelligent do they have to be?

If customers can order pizza, they don't need to be able to talk about the weather or the meaning of life. Their purpose is clearly defined and they need just as much intelligence to do the job well.

But here, too, there are different levels of intelligence. The simplest level can best be described with the Scheme "Q&A" describe. Digital assistants with this level of intelligence identify keywords and offer a certain number of suitable answer or choice options based on them.

But should they have a "guided process" If a product is able to master the concept of the customer, i.e. to conduct conversations with the conclusion of actions, it requires more intelligence. Products at this level must be able to recognise both customer intent and other entities.

The highest level of intelligence is the "Deep Reasoning"where conversations with complex content can be conducted. Each company can determine for itself the level of complexity that the product should be able to handle.

However, whether customers use digital assistants also depends on other factors. Not least on what customers expect from their counterpart - be it a conversation to pass the time, information on the weather or the stock market, or just a quick and easy pizza.



Our AT editorial team consists of various employees who prepare the corresponding blog articles with the greatest care and to the best of their knowledge and belief. Our experts from the respective fields regularly provide you with current contributions from the data science and AI sector. We hope you enjoy reading.

0 Kommentare

Submit a Comment

Your email address will not be published. Required fields are marked *