SAP has revealed that AI is now resolving one-in-five of its own internal support tickets without any human involvement.
Revealed during the company’s Q1 2026 earnings call, CEO Christian Klein confirmed that AI assists 100% of support cases across the company’s service and support function, with 20% resolved fully autonomously.
The deployment has delivered a 12% productivity gain, with teams handling higher ticket volumes without a proportional increase in headcount.
Klein told analysts on the call that “thanks to these efforts, we observed a 12% higher productivity in our support function.”
He also noted that SAP’s service and support teams are now “handling significantly higher ticket volumes without a proportional increase in headcount.”
Notably, this is the outcome that contact center leaders have consistently promised AI will deliver.
Reading the Numbers Carefully
The 20% autonomous resolution rate is the headline, but what surrounds itis equally important.
AI assistance across every single support case – not just straightforward ones – points to a model where agents are working alongside AI on all tickets, with the technology closing out a meaningful portion independently.
When combined with a 12% productivity gain and the ability to handle higher volumes without equivalent headcount growth, you have a picture that goes beyond headline efficiency claims.
SAP isn’t the first vendor to publish results from running its own AI across its support function.
Last year, Zoom reported that its Zoom Virtual Agent was resolving 97% of customer conversations without live agent involvement.
However, while there are clear comparisons to be made, Zoom’s metric reflects a self-service virtual agent handling largely standard product queries; whereas SAP’s 20% applies across a full enterprise support operation covering complex, often highly technical tickets.
Despite this difference, both examples do point to a growing pattern of vendors deploying and trusting their own AI in live support environments.
Another important thing to note is that SAP hasn’t publicly confirmed which products or platforms underpin these internal support capabilities.
We do not know, for instance, whether the support AI draws on SAP’s own service and CX portfolio, or reflects something purpose-built internally.
Part of a Bigger Internal Transformation
The support story sits within a wider set of internal efficiency moves made by SAP in the quarter.
The vendor has more than 80,000 employees in services, and Klein referenced consultants saving roughly one working day per week through AI-assisted system configuration and code analysis.
The company has set itself a target of around €2 billion in efficiency gains by the end of 2028, highlighting that the support function is one piece, not the whole picture.
Klein also touched on AI’s role in SAP’s go-to-market teams, citing over 83,000 hours saved and an additional €50 million of pipeline value from AI-personalized outbound campaigns.
Sapphire Is Worth Watching
For CX and contact center professionals, SAP Sapphire – taking place in Orlando next month – may well be worth keeping an eye on.
Indeed, Klein used much of the earnings call to tease what he described as “fundamental changes” to SAP’s portfolio, tied to a new agentic AI architecture, with the details reserved for the event itself.
“At Sapphire, we plan to announce some fundamental changes to our portfolio to infuse this deep domain know-how into SAP’s AI agents, and we will govern the agentic AI layer for our customers,” Klein told analysts.
“This foundational change will enable SAP AI agents at scale to deliver highly accurate results, to take actions across end-to-end processes in a secure manner.”
Product specifics were kept tightly under wraps. SAP’s portfolio covers CRM, service management, and customer engagement tools, so whether any of the Sapphire announcements land squarely in CX territory remains to be seen.
But given the internal support results Klein was happy to disclose in Q1, there will be no shortage of expectations heading into Orlando.
