Avantra CTO Says AI Must Prove Operational Value

Whether the AI investment boom holds is not a question SAP teams can answer. But Avantra Chief Technology Officer Jan Karstens says they can answer a more practical one: whether a specific AI application creates measurable, sustainable operational value.

In SAP operations, that value can be measured on the path from alert to recovery. AI earns its place when it helps teams turn operational signals into faster, more reliable responses. Avantra AIR is designed around that outcome-first approach. It analyzes SAP signals, surfaces likely causes, and helps teams move from an alert toward an informed response.

Start With the Operational Problem

In a recent article, Karstens wrote that enterprise AI deployments often fail because teams optimize for technical sophistication rather than operational usefulness. “The test for any AI application should be simple: does this solve a customer problem better than anything else available?”

He recommends beginning with persistent pain points: manual work that consumes excessive time, recurring issues that conventional tools have not resolved, as well as visibility gaps that create operational risk. Teams should define clear performance measures and require pilots to demonstrate improvement within weeks.

To illustrate, Karstens uses a critical SAP system anomaly. Traditional monitoring may generate an alert when a threshold is breached, but operators must still review data, correlate related events and determine the impact. AI creates measurable value when it reduces that investigative work and helps the team reach a reliable response sooner.

At the same time, he cautions against using AI when a simpler approach would work better. AI earns continued investment only when it solves a problem more effectively than existing tools, processes, and simpler alternatives.

Avantra AIR Must Meet the Same Standard

Karstens’ position sets a demanding standard for Avantra’s own product strategy. Avantra AIR, the company’s AI capability for SAP operations, must show that it can improve a defined process more effectively than the existing alternatives.

AIR analyzes SAP signals to explain an incident, identify likely causes and recommend a next step. It also surfaces the evidence behind its analysis, allowing operators to validate the findings before acting or escalating. That preserves human judgment while reducing the manual investigation that can delay resolution.

AIR can also connect its findings with IT service management platforms, including ServiceNow. This keeps diagnostic context inside systems where teams manage incidents, rather than requiring operators to transfer information between separate tools.

Avantra says early customers reported up to a 60% reduction in mean time to insight and resolution. They also reported a two- to threefold increase in automated tasks and monthly time savings of 10 to 20 hours per senior engineer. These results give customers defined outcomes to look for during a pilot: shorter investigations and faster movement.

Beyond response time, organizations need to assess how AI handles operational data.

AIR processes data within the EU, does not use customer data to train its models and deletes operational telemetry after seven days. The company says these controls allow AIR to analyze SAP incidents while keeping customer data private and under customer control.

Together, these criteria give teams a practical way to assess AI’s value: AIR earns its place only if customers can verify the gains in their own SAP environments.

What This Means for SAPinsiders

• AI value starts with a baseline. SAP teams need to know how long investigations take, where handoffs occur and which incidents consume the most specialist time. Without that starting point, faster resolution remains a claim rather than a measurable improvement.
• The alert-to-response process provides a practical test. Correlation, diagnosis, escalation, and judgment all happen within that interval. A credible deployment should reduce that effort without removing the evidence needed to make accountable decisions.
• Customer environments determine whether the gains are real. Reported improvements can shape a pilot, but SAP teams must reproduce them against their own baselines, workflows, and data controls. The value case becomes credible when faster responses improve decisions without weakening oversight.