Does a lot help a lot? No, at least not in app development. Because software doesn’t always have to be complex to offer a good user experience.
AI and machine learning are two omnipresent buzzwords. In many places people are looking for more or less meaningful application possibilities or at least talk about where potential lies dormant. But how necessary is artificial intelligence when it comes to a good user experience with apps , really?
How do apps actually work?
In order to answer this question, it is helpful to look behind the facade of the front end, i.e. the user interface. Apps are actually data collecting machines that fill a database – but nothing more: every click and every input is a specific command that, regardless of whether it is a mobile game or an app for language learning, is in the background in a record is transformed.
As long as this data is not processed further, nothing happens to it. So that something can emerge from this, they have to be processed according to a certain logic, for example counted. This is the prerequisite for being able to set different parts of the database in relation to one another, i.e. an algorithm is created that works according to the if / then principle, for example. Specifically: When the app is opened for the first time, information A appears; if the app has already been opened five times, information B.
So, based on the existing logic, firmly defined rules apply. The complexity can be increased here at will, with various dependencies and conditions, and possibly even continuously expanded to include feedback in the form of additional data. The crucial point at this point is that any processing of the available data by the algorithms must always be specified by a person and programmed accordingly in the system. Each individual usage path must have been thought out and created in advance.
So far, so understandable. If we now talk about AI and machine learning in the next step, then it is about the system working with the available data and the specified algorithms and independently recognizing patterns within the results that were not specified by humans.
Through appropriate training, the algorithm must learn which patterns it should look for. That’s the stuff scifi scripts are made of. For something like this to work, you need sufficient data, technical knowledge, training and tests – in short: a lot of resources. If they are actually available, wonderful application scenarios can arise. For example, in digital image recognition, in which the software recognizes better and better over time what is shown; this is particularly useful in e-commerce.
AI is also used in machine manufacturing, for example in the area of predictive maintenance, in order to prevent wear and tear on machines before an expensive breakdown occurs. If the resources required for the complex programming are not available, this results in a not fully developed solution, which in the worst case even has errors that lead to problems in operation. Programming without the use of AI is of course not child’s play either. The height of fall is higher due to the increased effort.
What is this excursion into the basics of app programming? It shows that it is not necessary to use high-end technology everywhere in order to offer a good user experience.
There is also “interaction” without AI!
To the outside world, and above all to the untrained eye, an app can appear to be intelligent, to learn over time or to interact with users. Most of the time this is not the case! A little clearer: If a lifestyle app is opened early in the morning, a “Good morning!” Appears and a message that a walk would be a good way to wake up. If the advice is not adopted, it disappears at a certain point.
It is the same when an app is designed for progress, as is the case with fitness apps or programs for learning a language. If the app is not used for a certain period of time, the goals will be adjusted because your fitness level or vocabulary may have changed.
In these two examples, the user gets the impression of an interaction because the app “reacts” to the individual use. However, the functionality is not based on any artificial intelligence, but on the in-depth knowledge of the app developers about the needs of the users and the correspondingly created and simple time- or count-based algorithms.
It makes sense to combine your own ideas from the development team with user feedback. In addition to explicit evaluations made in an app, feedback also means activity or inactivity in some areas of the app. This allows the user interest to be derived very precisely. So: to quickly test prototypes (for example in A / B tests), to improve them, and then to roll out the new functions widely. The result is an app that incorporates the needs of the user and – if desired – also includes interactive and responsive elements. This should show that a good user experience is not created solely by the technology used, but rather by making the best possible use of the resources available.
Balance between the possible and the sensible
When developing apps, it is always about finding a balance between technical finesse and user wishes. It can certainly make sense to adapt and implement new technologies in order not to lose touch with developments. The top priority should then always be the noticeable benefit for the user. From a purely technical point of view, for example, it is possible to map loyalty points via the blockchain. But is that noticeably easier for users in daily handling and does it improve the user experience positively? Of course it depends on the specifics, but probably not.
Instead of a technical sophistication that could not be implemented perfectly due to insufficient resources and then possibly not work properly, one should rather fall back on technically less complex, but functioning and well-developed systematics. Then apps that are “only” based on the natural intelligence of the developer teams can offer an outstanding user experience.