The Journey to Autonomous Support Through AI

We have all heard of “automated” support features, but as artificial intelligence (AI) has continued to advance, the support landscape is shifting to focus on the possibility of “autonomous” support. What exactly is the difference?

Here, Jens Trotzky, head of Artificial Intelligence Technology for SAP Support, discusses the role of AI in the development of autonomous support, and how SAP innovation is helping to make it a reality.

Q: What is the difference between automated and autonomous support?

A: Most customer support resources available today – even some of the most sophisticated – are considered automated in some capacity. In a nutshell, automated support is predictive technology that is pre-scripted by support engineers based on a fixed set of standards and defining factors. When we began exploring automated support at SAP, we started with a holistic, data-centric approach to deliver semi-automated recommendation systems that assist our customers and experts with incidents in real time. However, as enterprise software solutions continue to become increasingly complex, support requests will undoubtedly shift in terminology, urgency, and business context. In the case of automated support, the only way to ensure that these changes are recognized by support tools is for support engineers to manually adjust parameters on an ad-hoc basis, based on what they are observing from customer support-related incidents.

As the need for next-generation support continues to grow, autonomous enterprise software solutions support will undoubtedly become the standard across industries. Unlike automated support, autonomous support leverages machine learning and AI in a more advanced manner. Through permanent algorithm retraining, AI-driven support automatically adapts to changes in system environments or usage without manual input or adjustments. These self-improvement features enable support to become increasingly prescriptive to even the smallest nuances of the business and solution at hand, helping to meet the increased level of personalization that today’s B2B consumers expect.

Autonomous support will also be far better equipped to deliver proactive recommendations versus just a reactive approach. While automated support can auto-fill portions of a support ticket as a customer is typing out the issue at hand, an autonomous system will have identified the potential issue before it is even identified by the customer as an issue, and flag to the user to address it before issues arise.

What are some of the key benefits of autonomous support through AI?

Exponentiality is what makes the integration of AI-based technology, like machine learning, so transformative. Once a model is created, the possibility to scale becomes essentially endless. For example, as a business grows, so does its customer base and the need for customer support resources. Instead of having to multiply a company’s workforce of support engineers as customers are acquired, AI can accommodate anywhere from 10 to 10 million customers simultaneously through its algorithmic approach. And with machine learning taking each customer’s unique system preferences and goals into account, the delivery of predictive, proactive support is possible regardless of volume.

Click to enlarge

When businesses are able to run functions like customer support autonomously, valuable time and resources can be allocated to innovation-focused tasks that work toward value-add rather than troubleshooting. For example, by removing a majority of the manual work from support delivery, support experts can take on a new role where they can focus on standardizing processes and identifying usage patterns.

When will fully autonomous support through AI be realized?

We still have quite a way to go until fully autonomous support becomes a reality, but at SAP, we’re making progress every day. Currently, SAP Digital Business Services offers a semi-automatic solution: Incident Solution Matching via the SAP ONE Support Launchpad, which automatically connects users with the best-fit resources through natural language processing as they type. This kind of successful AI integration starts with analyzing customer-specific interaction data to identify which processes would benefit most from automation. Building on Incident Solution Matching, our team of data scientists have begun piloting SAP’s first fully-automated support service.

While I believe that we will someday be able to automatize specific aspects of the digital support process, I cannot envision a world where all B2B support is delivered autonomously. I’m sure that technological advancements will present this possibility in the future, but as studies show, 70 percent of consumers still prefer a human element in their support experience. As such, our goal in facilitating autonomous support is to elevate the role of the support technician by enabling monotonous tasks to run on autopilot with precision, while the engineers can focus their time and effort on innovating.

Overall, SAP’s vision for autonomous support is one that uses AI in a way that allows people – customers and support technicians alike – to adapt their roles at the pace of innovation.