Is intelligence enough for greatness or must it be paired with the ability to get things done and build collaborative action? This is a question that can be debated endlessly, but when it comes to artificial intelligence (AI) and machine learning, the answer is becoming clearer. For AI to be accepted and adopted, user experience must be taken into account.

In other words, great AI must be paired with a great user experience. So how do non-technical factors, such as design and acceptance, impact the efficacy of AI? How does the increasing use of cloud solutions change how AI is deployed? To understand the role of AI and machine learning today, it’s important to understand the context in which it is being employed – and the importance of context itself.

We sat down with Jens Trotzky, head of Artificial Intelligence Technology for Customer Solution Support and Innovation at SAP, to discuss the acceptance and application scenarios of AI-driven support. How can AI and machine learning-based features, solutions, and processes change the game and why is an AI mindset key for developing new solutions?

Q: Do you find that consumers are more attracted to AI than they were in previous years?

A: I would say there is strong interest and an uptick in customers approaching us to ask about our automation, AI, and machine learning capabilities. But I also think it’s a bit of a split topic when it comes to getting the actual end users to adopt the services that AI and machine learning provide. We still have to bridge the gap between the high-level executive view – which says “We want AI” – and the more day-to-day operations view, which asks how to make it actually work. There is a lot of education needed from our side for our customers, partners, and other stakeholders.

Additionally, I also see the expectation from customers that AI is somehow everywhere. Smart recommendations and smart helpers are expected. However, it can be a challenge to implement these functions so that they also achieve the targets we want. Of course, we want AI to add to a positive user experience, but we also want to improve processes for our users – to enhance them and speed them up. The challenge then becomes how to best integrate AI features in a way that both creates an attractive user experience and achieves these hard goals or KPIs.

Could you give an example of that?

Let’s take the support assistant. Support assistant is a guided workflow, in which a user creates a support request and is guided through the whole process. The user receives AI-based suggestions on the site and goes through an interactive dialogue. That sounds good and customers like it, but we also need to look at the numbers. What does this do to the number of tickets being created? We need to see the outcome of that process – it looks good on paper, but does it translate into tangible business results? For us, it turned out that a lot more process and AI optimization was needed. It can sound very nice to have AI in your processes, and customers want that, but at the end of the day it must also translate into tangible process optimizations for the customer.

What differs in how your team thinks about AI today?

“Context” is the key word, which increasingly moves us toward a world where every individual user can receive personalized support. For example, now we can determine what product the user is using. With the breakthrough of Built-In Support, we can capture the application context that caused the user problems and take that information into account. AI and machine learning allow us to tap into this vast knowledge of support history that we have about the customer. It helps us to connect the dots, like having some super support engineer on the SAP side that can magically sift through billions of pages of support documents that are out there and then pick up on the right cues at the right time to bring everything together. And with the help of the contextual window of Built-In Support, we can make all these connections in real time. And I mean literally real time. AI is a game changer for support because it allows us to know our customers, provide real-time insight, and help to mitigate a customer challenge before it becomes one.

Are you seeing a different type of user interacting with your AI these days?

Yes, for our on-premise software solutions the audience tends to be more technical – more subject matter experts themselves. With our software-as-a-service (SaaS) offerings in the cloud, we deal more with business users, rather than a technical person. We are, accordingly, trying to serve the right audience the right content.

This is exactly what we have in mind when we talk about Built-In Support because we realize that we are increasingly dealing with business users. The question is, what can we offer to business users? This is where integration comes in – realizing when and why a user is running into a problem and what happens next. Will they create a ticket? Do they want immediate support? Maybe it’s a simple how-to question. Then we need to offer whatever is appropriate: perhaps an answer using an AI-enhanced search algorithm, but maybe also offer a community solution. If, instead, we have a key user in front of that application, then we should offer the option of using things that might require key user credentials, such as a real-time chat session with an expert or the ability to create a ticket. This all happens because of context and personalization. In the end, we need to focus on the user out there and ask ourselves, what is their role? What is the task? What is the goal? I think the exciting part of support is shaping the interface between the user and the context.

There’s also the aspect of getting a unified user experience. In the end, the customer expectation is to have a single user experience anywhere in the SAP product portfolio. Here, Built-In Support can play a major role because by having the context, including history, the user experience can be truly unified. As a user, whenever I open Built-In Support, I can have the same experience, which can be customized or optimized for a specific product. The effort required to interact with support can be reduced and access to information can be faster than ever. I think that’s what most customers want: to help themselves.

What has been one of your biggest lessons from the last year?

I would highlight the role of integration. We have started to focus a lot more on the integration of AI and machine learning. We’ve learned the hard way how important it is to get the integration right. Often, it’s the integration that ultimately drives adoption.

Ultimately, the success of artificial intelligence is summed up in the equation user adoption multiplied by AI model accuracy equals realized business value. If you don’t have user adoption, then the value is limited. And often the biggest lever we have to drive user adoption is tightly tied to our ability to blend AI seamlessly into the user experience.

Sophia Stolze is part of Integrated Communications within Customer Solution Support and Innovation at SAP.