Ericsson builds SAP data fabric to scale AI across business

Ericsson is building a unified business data fabric with SAP Business Data Cloud as part of its push to expand AI use across the business.

The Swedish network equipment maker is creating a single governed data foundation for AI as it moves beyond pilot programmes into wider deployment. It is also progressing a transition to RISE with SAP and using SAP Business Technology Platform as part of a broader systems modernisation.

Ericsson operates mobile network infrastructure in 180 countries, and says more than 40% of the world's mobile traffic passes through its networks. That scale has made data governance and consistent business definitions central as the group looks to use AI in areas ranging from internal decision-making to operational processes.

At the centre of the plan is a federated data architecture that allows data to remain in existing locations while business semantics, governance and lifecycle policies are managed centrally. The approach is intended to cut duplication, simplify integration and apply the same definitions across SAP and non-SAP environments.

Ericsson says it has already moved from experimentation to wider use in some areas. More than 85,000 users are now live on unified Joule, SAP's AI assistant, according to the announcement.

Ericsson described its transformation as running on two tracks. One is the modernisation of core systems, including side-by-side extensions on SAP Business Technology Platform and a clean-core approach intended to reduce disruption to its ERP estate. The other focuses on using data and AI to improve decisions, raise efficiency and develop new sources of value.

Esra Kocatürk Norell, Vice President of customer experience, enterprise IT, at Ericsson, said the company's early investment in its data model was aimed at overcoming a common barrier in large AI programmes.

"Once you scale AI, it stops being an AI problem-and becomes a data problem," Norell said. "That's why we invested early in a business data fabric. With SAP Business Data Cloud, we can define what data means once-from revenue to market structures and access rules-and apply it consistently across the enterprise. That's what allows us to scale AI in a way that is trusted, repeatable and delivers real business value."

The tie-up also includes joint work on AI products. One example cited by the companies is an intelligent goal recommendation tool within SAP SuccessFactors that generates business-aligned goals for employees.

The tool is intended to reduce administrative work and improve execution, and is now being rolled out more broadly. The example offers a clearer picture of how Ericsson and SAP are trying to connect back-end data architecture with practical workplace applications.

Wider shift

The move reflects a broader trend among large companies seeking to shift AI projects from isolated use cases to standardised deployment across business functions. In many cases, the challenge has proved less about models and interfaces than about data quality, governance, access rules and consistency across legacy systems.

For telecoms groups in particular, the issue is sharpened by complex operating environments, large global workforces and a mix of old and new technology estates. Ericsson's emphasis on a federated architecture suggests it wants to avoid large-scale data migration while still imposing common standards across the business.

SAP said Ericsson's approach shows how customers are linking AI adoption more closely to business process design and governance. The software group has positioned its cloud data and AI tools as a way for large companies to build those controls into enterprise systems rather than layer AI on top of fragmented infrastructure.

"Ericsson's approach shows how leading companies are moving from AI experimentation to execution by focusing on data, governance and business context," said Manos Raptopoulos, global president of customer success for Europe, APAC, the Middle East and Africa at SAP. "Together, we are helping organizations unlock the full potential of AI at scale."