Artificial intelligence (AI) is one of the most talked about products out there. But what does it mean to us, to our product and our customers?
AI is polarizing. People are either very pessimistic or optimistic. It’s either all over-hyped, or it’ll solve world hunger.
Many researchers are looking at general intelligence – computers thinking like humans and handling many complex tasks and skills at the same time. It’s what the doomsday AI scenarios are based on in science fiction – but it’s a long way off still, and it’s certainly not much use in business yet.
Specific intelligence is what we can use now. It’s AI focus on a specific task to improve it. It manifests for us in three main ways:
- Human-machine interaction
- Predictive maintenance & service
Let’s look at each of them in turn.
1. Human/machine interaction
No, this is not Neo-jacked into the Matrix. It’s chatbots and natural language processing. As the chatbot is used more and more, it gets better and better at determining intent through machine learning.
For customers, Customer Engagement handles calls or other contact from customers rather than users through a chatbot in the first instance, which is really good at handling some queries. There’s need for human intervention or escalation if the chatbot can’t handle it – and that’s what our tool provides.
Internally, employees are looking for ways of interacting with their business applications in more natural ways. If you’re calling in sick, you can do so with the chatbot in Skype for Business on your phone, rather than having to call in or open your laptop, which is not a great experience if you’re lying on the sofa with a bucket by your side.
The next step is speech-only applications. If you’re driving, you need your hands on the wheel and eyes on the road, so you could only talk to your application. For limited, simple applications, it could work. And, indeed it does (de Vos demos it using Cortana on a smart speaker). A preview program is underway for IFS Applications 10 users, with Volac and Beijer Electronics Group as the first users.
2. Predictive maintenance & service
People have moved from calendar-based maintenance to usage- or condition-based maintenance, over time. The Internet of Things (IoT) has made this more common, with sensors flagging up problems. But AI opens up predictive maintenance. AI is good at finding patterns in complex data, that might predict failure states, based on historical evidence.
Adding AI to automate better is the next logical step in the evolution of automation. Business rules are a very useful form of automation – but they have limitations, including lack of variation over time, without development work. AI could help develop and shift those rules, initially by suggesting changes, but possibly, as quality increases, independently.
AI is looking at the historical data and determining what gets approved and what doesn’t. That information can be progressively used to automate more and more routine operations, as the system learns, freeing the user to do more valuable tasks.
Optimizations use AI to experiment with different scenarios with a speed humans couldn’t match, and predict the best outcomes. It can also react to problems and reassign tasks very quickly, by assessing all the possible options and picking the best.
A digital twin of an object is a common idea. NASA has been thinking about it since the 60s! However, it’s taking on a new significance in the current era
- Anytime anywhere – our digital twins can be accessed anywhere
- All Assets – anything can be modeled and monitored through sensors
- Advanced analytics – allow you to monitor and learn quickly
- Beyond the asset – we can understand the relationship between the digital twin of the asset and the digital twin of the business
For example, you can play back the moment a robot arm went wrong, and through the animation of its digital twin, understand what caused it to malfunction. That’s information you can use to prevent the same things occurring again.