5 min Applications

IFS wants to apply explainable AI to customer data by 2020

IFS wants to apply explainable AI to customer data by 2020

During the IFS World Conference, we had the opportunity to look ahead with IFS directors. Where does the company want to go in 2020? It sees itself as a challenger, but how will it really challenge the competition? IFS is also entering the world of artificial intelligence (AI). The company wants to make the application of AI in 2020 concrete for its customers as soon as possible.

IFS has been paying close attention over the past period. The way in which the company unfolds its plans for artificial intelligence (AI) is just a little different. The average CIO and IT professional is pretty enthusiastic about AI. Everyone does something with AI these days, but the customer still sees far too little of it. At IFS, they seem to realise that.

At IFS, like the rest of the market, they are convinced that AI is ready for use. With AI, you can’t automate the entire ERP solution or make it redundant, but you can solve business problems or gain better insights. To do this, you need a good data lake.

So far, that’s not much new information. IFS does, however, make a difference in this area. IFS’ ERP solution should soon be able to automatically pour all the data from the ERP system into a machine learning model so that the customer has direct access to the models and AI solutions. Above all, the new technology must be easy to use. To ensure that the models and the AI are powerful, customer data is used to generate the models. This means that no generic models are released on the customer data, which would increase the chance of errors and deviations. Every company and every market is different, and with their own model the results are better.

IFS also chooses to make the AI fully explainable. This way, the customer can always trace how the technology came to a certain outcome. If the entire process is clear why there is a certain outcome or advice, it can also be quickly estimated whether there are errors or prejudices in it.

IFS is working on a generic machine learning and AI platform that will soon be easy to implement in the company’s software. Ultimately, it simply wants to provide quick and easy access to new technology. Don’t dwell too long on all the theories and possible use cases, but put the technology in the hands of the customer. A strategy that we can appreciate. AI is often far too theoretical.

The platform that IFS develops is partly based on open-source standards and models, but also partly on proprietary technology. Within IFS, AI has been used for some time now. In the software of IFS, AI has already been applied on several fronts to improve products, e.g. to make business processes more efficient. This can involve tackling actions that are often carried out or improving the deployment of assets within the company.

Cheerpack robots controlled by AI

Packaging company Cheerpack uses robots in its warehouse to pick orders. This is done fully automatically and is controlled by the IFS software. As soon as new stock is needed somewhere, IFS will send a signal to the robot platform to request that new stock should be picked from a specific location and delivered to location Y. The robot software then ensures that the right robot completes that task as quickly as possible. The intelligence lies in the fact that IFS can keep track of when which product is needed where and submits a new robot order on time. To achieve this, machine learning and AI are required.

Another frequently heard example is the timely maintenance of machines. If certain sensors measure that the performance of large material is decreasing, this material reaches the point where it is wise to do maintenance before the material comes to a standstill due to a malfunction. IFS has already developed the technology for this.

IFS believes in RPA as an extension of AI and machine learning

Choosing the right model and having the right technology to connect different systems remains the hardest part. Since IFS is active in many traditional markets where outdated solutions are often used, it is not always possible to directly link up with other technology. There are simply no APIs available. Therefore, IFS sees many opportunities through the use of Robotic Process Automation (RPA).

What RPA actually does is operate an interface like a human being. It can fill in forms logically and press the appropriate buttons, depending on the data input. APIs may be faster and simpler, but if that technology is not available, RPA can offer an affordable solution. Replacing existing machines and technology costs millions of euros in some industries.

IFS can make it happen

We talk to a lot of companies all year round, who all want to talk about AI when they get the chance. Often it remains very theoretical, and the conclusion then is that everything will be possible within a few years. On the other hand, there are companies that simply apply AI and machine learning, so that you, as a customer, can immediately get to work with it. This group of companies is in the minority for the time being. IFS seems to be joining that minority in 2020. We are very curious to see how that will play out. In any case, for the customers of IFS, it is a welcome addition.

IFS has shown that they can change and innovate a lot in a short period of time, read our article about it: IFS moves from slow enterprise to fast innovator