Today, it can seem like everything — and everyone — claims to use artificial intelligence (AI) or machine learning. But what does that really mean and, more specifically, how can it be used to help support SAP customers?

There are plenty of outcomes that AI could be used for, but how do we use it to maximize the impact and how do we connect it to businesses outcomes? And perhaps most importantly, how do we develop an AI mindset that will help us innovate the uses of AI in different products and for different businesses?

It is clear that we have only begun to realize the potential of this technology, but it can be hard to understand where it should be applied next.

We sat down with Jens Trotzky, head of Artificial Intelligence Technology for Customer Solution Support & Innovation at SAP, to discuss how SAP’s automated support strategy is changing, what an AI mindset looks like, and what to look for next.

Q: How do you see SAP’s AI strategy evolving this year?

A: We started off with something I like to call semi-automated machine learning, in which we try to recommend things to customers or internal stakeholders, but a human decides what content makes it to our customers or the customer themselves decide whether they want to accept what the AI suggests.

Now, we are heavily investing in predictive support, which is increasingly called preventative support. In this model, rather than having an event take place and then having the AI provide possible reactions, we try to anticipate the event itself and, when appropriate, prevent it. This is part of an effort to help our customers benefit from one another. Previously, if one customer had an issue, other customers wouldn’t benefit from their experience, at least not anytime soon but now through AI-enabled predictive support, that’s changing.

And one other part of the strategy worth noting is an increased focus on personalization. We are moving away from a one-size-fits-all strategy towards a more targeted approach. That’s possible because we’re curating more and more data from our services and support business. That allows us to then provide a more accurate, clear context around each of our customers, and then support them accordingly with machine learning.

How have you seen the “AI mindset” changing within SAP and among customers?

Previously, we would have other teams within SAP coming to us saying, “I have a business problem that I want to solve. How about we just use some AI?” That indicated we needed to develop that AI mindset within SAP, because AI is fundamentally a data-driven topic, and so must be approached from that perspective. It’s the data that tells the story; it’s up to the company to unearth the potential that lies hidden in that data.

This can be even more of a challenge outside of SAP, because businesses are focused on their business problem first and foremost. Of course, we want to be inspired by the business challenges that we face, but the solutions ultimately grow from within the data. I think there is a growing understanding that you must pair a business requirement with data science. And then be prepared to make changes to your business process to collect data that you currently don’t gather to make it useful in AI processes.

One great example of the better understanding of the uses of AI among our external customers is a change that was made to SAP Ariba Customer Support using an Incident Solution Matching service. This service attempts to solve a problem by suggesting a solution before resorting to creating a support ticket, which customers really like. Last year, SAP Ariba support started combining a classical search technique with AI. In brief, after the customer asked a question, the AI would guide the customer to provide more information. AI was deployed to refine the search by asking additional questions. By using AI in this way, the technology was brought into the customer’s support experience at the right point to have the maximum impact on getting the customer the needed support quickly and easily. We want to ask not just whether we can use AI for this, but how and when we introduce AI to have the most impact.

What do you see happening with AI in support over the next year or two and what are you most excited about?

One of the things that we are working on is called “process intelligence.” Instead of thinking of everything as being a single transaction that we want to accelerate with AI or machine learning, we instead look at the overall customer journey. By mapping out that journey, we want to make use of these technologies to increasingly predict the path the customer would take and use AI and machine learning as an agent in the process, helping the customer journey along.

Next, we will speak with with Jens about how user experience has influenced the use of AI in customer support in recent years.