Software of the future is not a tool for work – it is work itself!

A few years ago, I had the task of carrying out an evaluation of various ERP systems with a focus on the financial sector. It was not the first time I had done this. I had quickly put together a good list of requirements, talked to various people who were much more deeply involved in the subject than I was, and organized offers, information and demos from providers. I was confident that I would be able to make an informed and good decision.

What does software cost?

The costs of the various providers were also quickly compiled. I spent a little time running through various processes in the respective applications. I was amazed at how differently time-consuming it was in different products to operate one and the same business process with the functionality provided. What was done quickly in one application required considerably more operating time in the other. That made me wonder.

At the end of the day, as an entrepreneur I am primarily interested in what a certain process costs in total. If my employees have to spend a lot of time operating software, the process becomes much more expensive.

So I went and benchmarked the various products against the requirement “time consumption in key processes” and was astonished to see that the picture of the various providers had changed completely. The cheapest provider in terms of license and maintenance costs was suddenly – after considering the total costs – in the bottom third.

The time differences were sometimes really enormous. There were differences of a good 800% – for one and the same task, nota bene.

Usability

User-friendliness, which is reflected in a significant reduction in the “operating time” of the software, is therefore one of the killer criteria par excellence. Despite this, it is still neglected (a quick search revealed that very few people take operating time into account during evaluation). In fact, the time that employees spend using software is in most cases the biggest cost factor.

This made me realize that, at the end of the day, usability should be measured in costs. The formula for this is simple:

Some usability experts will now comment that this is a profane reduction of the whole usability topic to costs. That usability is so much more. On the one hand, this is probably true – on the other hand, colleagues in the usability field are often accused of operating in a rather undefined area that is difficult to measure. I think that with this definition of usability, we have created a tangible and relevant metric for business software that can be used to compare different solutions.

Artificial intelligence

One of the central challenges of software product design is therefore the reduction of operating and interaction time. On the one hand, this can be achieved by radically focusing on a single segment and/or a single use case (cutting out features). On the other hand, it can be achieved by reducing the complexity of human interaction with the software using artificial intelligence. My simple idea at the time was that you could simply let the system learn how it is operated by humans so that it can take over the operation itself over time.

Since this realization, this narrative has stayed with me. Because if a system can serve itself over time, this also means that it can take over the actual work of humans itself.

And it also means that we can transform an intuitive and natural conversion path for humans from software – as we know it today – into autonomous, intelligent software. My two start-ups Parashift and Accounto are, among other things, a product of this thinking and conviction.

AI: hyped and underestimated in equal measure

Artificial intelligence is being hyped today. So-called tech influencers in particular are spreading urban legends at a rapid pace via social media. They often remind me of skateboard videos in which a daring rider performs a super complicated trick over a huge staircase. What you don’t see is that it literally took 250 attempts to pull off this one trick once. It’s the same with the robots that heave parcels around or jog through rough terrain. What we don’t see are the thousands of failures.

What is underestimated, however, is the paradigm shift that artificial intelligence will bring about in the software industry. I find it astonishing that the software industry in particular, which still relies completely on manual work in very large parts, is only very slightly aware of such impending changes.

From tool to digital service provider

In the end, this change will transform software from the tool it is today to facilitate work into a service provider that can do work autonomously. You can call it a robot or whatever you like. I believe, and we are working on this in our teams, that sooner or later software will become a service provider almost unnoticed.

The economics of lower personnel costs in operation pave the way. In the end, if you like, artificial intelligence will create usability that is simply unbeatable before there is no longer any need for interaction with autonomous software.

This article originally appeared on the PEAX blog