Artificial Intelligence (AI) is such a big topic, and so many different possibilities exist, that you should focus on those opportunities that make the most sense for your business. You can’t do everything at once and that applies to us, the Research & Development teams of IFS, as well. Which areas make sense? Where can we add the most value to our products by leveraging AI technologies?
In my earlier blogs, I already established that general or human intelligence is quite far away for machines. And although this might be the future of AI, we’re not close yet. This is not AI as it can help us in our business today. At IFS, we like to keep things pragmatic and practical. That means we focus on something I call specific intelligence. We’re still training computers to do things smarter than we might be able to do ourselves, but it’s often a very narrow, very specific use case.
AI for us is a means to a goal and not the goal itself. It is about using technology to make better solutions for our customers. I believe that with AI technology we can empower our users to do more with less, allowing them to be more efficient and effective in the jobs they do, freeing them up to add more value to their organization.
That means we decided to focus our AI efforts in three areas because we believe that’s where it would contribute most in providing better applications to our customers.
- Human-machine interaction
- Predictive maintenance and service
Let’s have a look at these three areas in a bit more detail.
1. Human-machine interaction
Again, we’re not talking about Neo who gets jacked into the matrix. It’s about how we can improve the day to day usage of a business application to enhance the user experience. There are many areas where we can do this by leveraging technologies like chatbots or computer image processing.
Let’s take chatbots inside of IFS products as an example. There are two ways to look at the usage of chatbots in context of IFS products. The big difference is whether we take the viewpoint of an external user, for example, a customer that is contacting you, or whether it is an internal user, for example, an employee, that wants to have easier access to the business software.
As we differentiate between these two use cases, we have introduced two different solutions in our portfolio: IFS Customer Engagement, focusing on customers contacting your business, and the IFS Aurena Bot, focusing on providing a more conversational user interface for IFS users.
2. Predictive maintenance and service
With the rise of big data and advanced analytics, an area with a high pace of change is predicting and forecasting. There are plenty of compelling use cases, too. For example, many years ago I ran a project for a customer to predict the need for an increase in back-office workforce based on the effect of marketing campaigns.
However, if we as IFS want to focus and if we look at where we feel our customers are benefited the most in the short term, it is about predictive maintenance and predictive service. This is about using IoT-generated data and machine learning to better predict failures and maintenance or service needs.
Traditionally many assets have been maintained based on calendar-based maintenance schedules. Some industries have been more advanced and moved to usage- or condition-based maintenance long before. For example, aircraft engines are maintained based on flight hours or engine cycles for many years. Today, with the rise of IoT, this is becoming reality for many industries. As IoT sensors are collecting lots of data on the asset, we now have the opportunity to move to predictive maintenance and service as well.
The IFS IoT Business Connector provides the machine learning capabilities to develop custom algorithms and models for specific use cases or IFS provided solutions. In addition, it contains easy integration points to connect to third-party discovery tools, often provided by the OEMs of an asset.
The third and perhaps the most interesting area is automation. We started to develop business applications to help automation processes. Although much is automated today, many decisions are made manually, and that’s fine, but often there is room for automating more by implementing business rules. For example, if the purchase order value is smaller than $50 it can automatically be approved.
But business rules have shortcomings. Firstly, there is a limitation in complexity, secondly, business rules don’t dynamically change over time. Therefore, the next logical step in the evolution of automation is adding AI to capture that complexity and dynamic in a self-learning model.
In this area of intelligent automation, we’ll focus on three types of automation:
i. Anomaly detection
Anomaly detection is about alerting the user and drawing attention to patterns that the user would most likely not have looked at otherwise. For example, a top five list of invoices most likely not be paid on-time or a top 10 list of work orders most at risk of not being completed in time. The benefit being in the handling of large amounts of data and being able to find patterns that humans are unable to see.
ii. Single transaction
Then, the automation of a single individual decision or transaction is the next step. With help of, for example, pattern detection algorithms, we can start automating decisions based on the likelihood they’ll be correct. Although this will often start with recommending a decision, with growing quality and confidence in algorithms, mass automation could be in the future.
Take expense approvals for example. The moment an employee commits his expense report, a machine learning model would be used to predict the likelihood that the expense report is correct and automate the approval or escalate to the manager. That way we would allow the user to focus on the things where he can add value as a human instead of spending time on highly repetitive work.
Optimization is taking a complex situation with many variables and input and creating the optimal strategy for making many decisions. Optimization is seen in many areas, ranging from supply chain to manufacturing to field service.
A perfect example is how IFS Planning, Scheduling and Optimization has been using AI capabilities like neural networks for the past 12 years in order to create the most optimal schedule for field workers.
To express how complex this can get, an average field service schedule has more possible outcomes (10179) than chess (10128) or go (10170), while our largest customers can go up to 100,000 activities and 10,000 resources, which is impossible for humans to plan optimally.
What AI means to IFS
In this blog post, we have explained what AI means to IFS products and our strategic direction forward. It’s about focusing on human-machine interaction, about predictive maintenance and service and about automation. But even more importantly, it’s about augmenting our customers, our users to be better equipped in the jobs they do, for them to be even more efficient and effective tomorrow than they are today—and to do this in the way the world is used to from us; pragmatic and practical, based on real-world customer cases.
Learn more about IFS Labs at ifs.com.
Do you have questions or comments about IFS’s AI investments?
We’d love to hear them so please leave us a message below.