License to skill

Polling the weather with a voice command, switching on the lights and the heating system and placing orders in the “biggest supermarket in the world” saves the grasp on the smartphone. Although the first industry reports show that voice shopping is not yet receiving the acceptance that Amazon expected from the combination of its core services.

You want to know more about voice marketing? Day three of the Munich Marketing Week is all about Alexa, Siri, Cortana and Co.

However, Voice User Interfaces (VUI) like Alexa, Siri or Google Assistant will become an integral part of the game as they provide a very comfortable and fast way to interact with services and brands. For companies and hobbyists Amazon offers the opportunity to offer their services through the so-called “skills”. These expand Alexa’s repertoire by the ability to call up the services of a provider by voice. In 2018, more than 50,000 skills existed from more than 3,500 brands.


There is a skill for everything

According to Apple’s trademarked “There’s An App For That” trademark, Amazon may soon take up the cause of having a skill in hand for every application. Starting with its own services, the Alexa device Echo plays music on Spotify or Amazon’s own streaming service, launches movies and TV shows on the Fire TV stick – including the largest video-on-demand competitor Netflix.

Here Amazon delivers a remote control with microphone, which supports Alexa. In addition to polling the weather and setting alarms and timers, Alexa provides a rudimentary to-do list, reading messages from preset sources, and answering questions you would otherwise google.

However, the third-party skills are what make the Alexa devices Echo, Echo Dot, Echo Show with additional screen or echo car so powerful for the vehicle and which should have been particularly pleasing to smarthome manufacturers. Instead of pulling the smartphone to control each connected light bulb or heater, Alexa combines all the skills under one hood and moves the operation to the natural language level. The future scenario of the automated and intelligent house of old science-fiction films became reality almost overnight and should have contributed significantly to the success of language assistants, including that of Alexa’s biggest competitor, Google Assistant.

Jan Antonin Kolar 1048698 Unsplash
Photo by Jan Antonin Kolar on Unsplash

In the core business of Amazon, however, after initial euphoria showed that customers hardly order by voice and if, then usually only once and never again. A study of The information According to estimates, only one in every 50 has placed an order with Alexa at Amazon and nine out of ten have never done so again or have later canceled the ordering process. Further studies show similar weak results. One can speculate why customers do not order via Alexa even for smaller, everyday purchases. The Neckermann or Rewe skill offers here, for example, no purchase function, but merely informs about offers.

Ironically, however, customers are using skills to order food, such as those from Lieferando or from Foodora, which only gives the arrival time. As is so often the acceptance in the USA is higher here than in Germany, orders with the E-Commerce giant seem to have hit however still no nerve world-wide. And that although Amazon bids algorithms and offers well-rated products (marked “Amazon’s Choice”) when a customer does not order a specific brand, but just a product. Even discounts on merchandise ordered via Alexa do not seem to be the key to success yet.

Who celebrates, can also work

Anyone who is tempted to think that their webshop might not initially be considered a skill could possibly succeed with other applications for their company. When it comes to non-consumptive services, Alexa is in fact more productive than doing a game on the Xbox. To get up from the couch, for example, there is a skill variant of the popular 7-minute workouts, which lashes verbally as a personal coach through the exercises. The Fitbit skill of the world’s largest provider of fitness bracelets will then directly through the burned calories.

Also interesting: the sound of the future

Alexa skills can also be used to train the brain muscles, for example to learn languages ​​- in other words, what the AI ​​itself is constantly doing in the cloud. Meanwhile, there are several such skills, ranging from vocabulary coaches, over dictionaries to translators from Swabian to High German. If the applications are extended to other areas of daily life, Amazon has become active itself and wants to bring the language assistant to the office and conference rooms of the world with “Alexa for Business”.

For customers of its own and world’s largest Cloud Amazon Web Services (AWS), the Seattle-based company offers a subscription model that includes a set of skills designed specifically for everyday work. The booking of conference rooms with the corresponding entry in the company calendar can be done just as well by voice command as the start of telephone and video conferences or the request of information from tools of Salesforce or ServiceNow. For the latter case, Amazon provides an interface similar to normal skills to pass voice commands to each tool and database. And it is exactly this interface that basically anyone can use to build their own skill. Even without programming knowledge.

How do skills work?

The recognition of natural language has not been a trivial act for years and only the development in recent years has made it possible for language assistants like Siri or Alexa in the first place. Amazon bundles its recognition software and a set of algorithms in Alexa Voice Service (AVS) and makes them better and better with machine learning. Basically, the same thing happens with every request: Alexa waits for the hotword to activate, then records the command, sends it to Amazon’s cloud for evaluation, executes the appropriate program code of the skill, and provides confirmation on the echo or other device integrated Alexa out.

If the AI ​​in the cloud does not understand the command because the user repeats a similar-sounding command immediately afterwards, machine-learning algorithms are used that make speech recognition more and more accurate. What privacy advocates often bellow, will ensure in the future that Alexas speech recognition will eventually be so good that they can even the lamps on a command with a strong accent in the drunken state can approach.

Incidentally, developers of skills do not have to deal with this natural language processing (NLP) and the technology behind it. AVS takes over this part completely and hands over the heard to the skills. For their development, Amazon offers a number of skill kits, so templates that are equipped with certain features and optimized.

It can be used to create skills that play music or videos, control smarthome devices, read messages or generate lists. A custom skill set, which can accept and process any kind of request, offers Amazon additionally, here of course more programming is necessary.

Skill development according to modular principle

If the nature of the skill is fixed, then it goes to the most interesting part of the development, namely the preparation of the interactions. Amazon distinguishes between several interaction models: From the query itself via queries by Alexa to the user’s response and the optional execution as well as the verbal confirmation or an answer to the user’s request, for example according to the weather.

Basically, the right workflows just have to be planned and then transferred to Amazon’s AWS-Lambda service, whereby another cloud provider can be used here as well. Some code and web space for optional images, audio and video files and of course an Alexa device for testing are actually all it takes to create a skill.

Those who do not speak AWS lambda-supported programming language can also experiment with Amazon’s skill blueprints. These templates make it easy to create mini-games or smaller skills, such as a personal trainer or a virtual assistant, who can provide information to roommates when they are away. If it then becomes a professional skill for a company, tools are available on the Internet with which one can create skill prototypes literally with just a few clicks.

Invocable is one such tool that provides a clean user interface that not only allows the generation of language skills, but also those that display videos or images on the Echo Show and third-party devices. Botsociety also allows you to build prototypes of Google Assistant applications as well as chat and messenger bots for Whatsapp, Facebook Messenger and others. Conversation.One and the tool Voice Apps are a little less fancy, but offer free subscription models, which can be extended depending on the application for a monthly fee to features. Considering to develop several skills or bots and services for other platforms, such modular solutions are a good tool, if it is to go fast and easy or without a programmer.

What is the purpose of a skill?

Not insignificant should be the question, which skill offers itself for the own business, which can be answered with a clear “it depends on it”. As with the development of apps (see LEAD 03/18), a skill should not be developed because it is the next big thing, but it should make sense to expand your own product or service.

A good example of this is Uber, whose skill allows calling a vehicle. Here, the chauffeur service provider has simply implemented the interface of the app as a voice user interface and thus created real added value. Also conceivable are skills that can be used to reserve a table in a restaurant or make an appointment with a plumber or craftsman from the immediate vicinity. Basically, it is enough to ask yourself the question of removing a hurdle from your (potential) customers and using the skill for a mutual purpose. From the perspective of the skill developer, of course, in the form of increasing customer loyalty and sales or as an extended communication channel – certainly also in the area of ​​content marketing.

No matter what the business model looks like, it is best to analyze the skill usage. Amazon offers a visually prepared interface with charts and tables, which can probably be summarized as the easiest way Skill Analytics. Here you can see anonymously how, when and how often the users use the skill and where there is still a need for optimization. Depending on the application, analysis of the data may be used in addition to tailoring other business areas that are not directly related to the features of the skill. Basically, the possibilities here are varied and likely to be as sophisticated in the coming months and years as SEO tools or apps today are.

The competition is silent

Amazon was with Alexa, although the provider that made language assistant socially acceptable, but in the future against Mr. Internet will have to say: Google. Alphabet’s driving force has written more than a year ago artificial intelligence and deep and machine learning on the flags and pursues this goal with its own hardware and integration in all Android smartphones with iron will.

Amazon continues to hold around 70 percent of the market share for voice assistants, but Google’s 25 percent could grow significantly thanks to data from heterogeneous sources. Apple’s HomePod with Siri does not really matter at around five percent, and Microsoft’s Cortana and Samsung’s Bixby will probably soon disappear into oblivion, as both companies just do not have enough data levers to do that.

What is certain is that the market of skills and providers will continue to grow and more and more sophisticated services will come.

Even in Germany, the language assistants are becoming increasingly popular and can be found in almost every household in a few years, if they are not already being used on the smartphone in the form of Google Assistants. It remains to be seen whether we will be in for a renaissance like the advent of apps, and whether purely business-based business models will emerge. This will probably also depend on how Amazon expands its interfaces and what Google contributes to competition to revive the business.

Also interesting: Siri and Alexa marry

Newsletter & Messenger

Always up to date on all topics of digital life with the LEAD Newsletter and the LEAD Tech Newsletter. Whether professional or private. In your inbox or via messenger.

Subscribe to our newsletter now
Subscribe now via messenger

Leave a Reply

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