Mobile World Congress in Barcelona remains one of the telecommunications industry’s defining annual gatherings, and the themes shift each year — from smartphones and successive generations of mobile networks to cloud, IoT, edge, and now AI. What stays constant is the central role connectivity plays in everyday life: people simply expect calls, video meetings, navigation, streaming, healthcare access, and business operations to work. Behind that simplicity sits a vast, complex web of infrastructure, and ahead of MWC 2026 Databricks framed how data and AI are reshaping how that infrastructure is run.
Telecommunications at a crossroads
The industry faces a familiar squeeze: rising demand and capital costs on one side, and pressure to operate more efficiently on the other. Critical decisions — where to invest in the network, which customers to prioritise, how to optimise field operations — often cannot be made quickly or confidently enough, because the underlying information is hard to access and trust in real time.
Enterprise intelligence requires data intelligence
At the root of these challenges is a common problem: data that is fragmented, slow to access, and difficult for teams to rely on. The argument Databricks advances is that data strategy and business strategy are now inseparable, and that improving financial performance increasingly depends on becoming a genuinely data-driven organization — one where decisions about marketing, network investment, customer service, and fraud prevention are guided by reliable, near-real-time insight. The framing is deliberately ambitious: not more dashboards, but a different operating model in which accurate data flows continuously, AI keeps learning from it, and operators shift from reacting to events toward anticipating them.
From data intelligence to measurable results
Fraud prevention is offered as a concrete example. Historically, investigations were largely manual, slow, and costly. With integrated data and AI, the described approach identifies anomalies in real time, extracts fraud signals from unstructured interactions such as call and chat transcripts, and models likely future fraud patterns. Operators can, in principle, monitor large customer bases at once, improving detection accuracy while reducing false positives, automating checks that once required manual effort, and deploying new models far faster. Databricks cites customer examples reporting large reductions in fraud attempts — in some cases on the order of 80% — and substantial annual savings, figures that come from the vendor and specific deployments rather than independent benchmarks.
What it means for telecom leaders
The implied playbook for operators is to anticipate and resolve network problems before customers notice, optimise and personalise customer channels in real time, continuously tune operations, guide investment by actual usage patterns, and reinvest savings into growth. The broader claim is that operators will increasingly be judged not by coverage or speed alone but by how intelligently they use their data to serve customers and industries.
Limitations and what to watch
This is a vendor perspective tied to a marketing moment, so the performance figures — particularly the fraud-reduction numbers — should be read as illustrative results from selected deployments, not guaranteed outcomes. Moving to a continuous, AI-driven operating model is a significant organizational change that depends on data governance, integration, and skills as much as on any single platform, and real-time fraud and customer systems raise privacy and regulatory obligations that vary by market. The durable point underneath the messaging is sound: telecom decisions improve when reliable data is available quickly, regardless of which tools deliver it.
The original commentary appears on the Databricks blog. For related reading on this site, see coverage of real-time data processing and building applications on Databricks.
