Telecommunications operators are eager to leverage artificial intelligence to drive innovation and operational efficiency, but many still find themselves stuck in the pilot phase. Despite increased investment in AI and notable progress in customer-facing areas like marketing and customer care, telcos consistently struggle to scale AI adoption across more complex, network-centric domains. Key obstacles include fragmented data architectures, limited integration between systems, and the absence of unified, AI-ready infrastructure—challenges that often result in promising pilots failing to deliver enterprise-wide transformation.
To address these hurdles, the industry is witnessing a shift toward hybrid data cloud platforms. These architectures offer telcos the flexibility to run AI workloads on-premises, at the edge, or in the cloud—wherever it makes the most sense for data sovereignty, latency, and cost optimization. By enabling models to operate closer to the data, hybrid data cloud solutions allow operators to overcome silos, accelerate real-world AI impact, and realize tangible business value while maintaining regulatory compliance and performance at scale. This strategic pivot is key for telcos seeking to unlock the full value of their data and transition from isolated AI pilots to organization-wide transformation.
From AI pilots to scalable impact: How telcos can unlock value with hybrid data cloud platforms
Artificial intelligence is firmly on the agenda for telecommunications operators. Yet despite growing investment and experimentation, many telcos are still struggling to scale AI beyond isolated use cases. In a recent Analysys Mason podcast, Adora Okeleke, Principal Analyst at Analysys Mason, spoke with Laurent Laisney, Telecom Industry Strategist at Teradata, about the current state of AI adoption in telecoms — and what leaders must do to turn promise into performance.
AI adoption is accelerating — But uneven
AI adoption across the telecom industry is clearly gaining momentum. Operators have made notable progress in customer-facing domains such as marketing, customer care, and other Business Support Systems (BSS). These areas often offer faster returns and clearer business outcomes, making them natural entry points for AI initiatives.
However, adoption is relatively less mature in network-facing and operations support systems (OSS) environments. AI use cases in areas such as network optimization or anomaly detection are often narrow in scope, siloed, and difficult to scale across the full network lifecycle. As a result, the industry is seeing an imbalance: innovation in pockets but limited end-to-end transformation.
Listen to the podcast
Listen to the full podcast episode to explore expert insights on AI adoption in telecoms, the challenges operators face today, and how hybrid data cloud platforms can enable scalable, real-world AI impact.
What’s holding telcos back?
Several structural challenges continue to slow down AI deployment at scale:
- Fragmented data architectures: Data is often scattered across OSS and BSS silos, making it difficult to train, deploy, and govern AI models consistently.
- Limited integration: AI capabilities embedded in vendor-specific solutions rarely interoperate across domains, restricting scalability and performance.
- Lack of unified AI-ready infrastructure: Many operators lack platforms that support real-time inference, lifecycle management, and explainability across hybrid environments.
- Unclear ROI: AI initiatives are frequently exploratory, making it hard for executives to connect pilots with tangible business value.
- Skills and cultural barriers: Shortages in AI talent and resistance to organizational change often prevent AI from being embedded into core workflows.
The result is a familiar pattern: promising pilots that fail to transition into production at scale.
Why the hybrid data cloud is gaining traction
Looking ahead, the conversation highlighted a clear shift in how operators are addressing these challenges. Increasingly, telcos are investing in hybrid data cloud architectures that allow AI workloads to run where they make the most sense — on-premises, at the edge, or in the public cloud.
This approach reflects the realities of the telecom environment:
- Data sovereignty and compliance require sensitive network data to remain under operator control.
- Latency and performance demands make public-cloud-only models impractical for real-time AI use cases
- Cost management is critical, as unpredictable public cloud costs and inefficient GPU usage can quickly erode returns. By bringing AI models to the data — rather than moving large volumes of data around — operators can improve performance, reduce risk, and maintain regulatory compliance.
Key considerations for business leaders
For executives evaluating hybrid data cloud strategies, five considerations stand out:
- Data sovereignty and compliance: Ensuring sensitive data remains protected in regulated environments.
- Latency and performance: Supporting real-time AI workloads with low-latency access to data.
- Scalability and cost control: Balancing cloud elasticity with predictable economics.
- Integration and interoperability: Enabling seamless data flows across OSS, BSS, and operational systems.
- AI readiness: Supporting full AI lifecycle management, transparency, and explainability.
AI success is no longer about algorithms alone. It’s about strong data foundations, integrated architectures, and relentless focus on business outcomes.
Agentic AI and the Autonomous Enterprise
According to Teradata’s perspective, platforms designed for portability and consistency across environments help operators avoid vendor lock-in, maximize existing investments, and respond quickly to changing business needs.
The future of enterprise architecture is intelligent, adaptive, and designed to think. Teradata’s ambition is to be the leading AI and Knowledge Platform for hybrid environments — where data spans cloud and on-prem systems.
The next phase of AI is not just predictive or generative — it’s agentic. Autonomous agents will reason, plan, and act across enterprise systems.
Teradata is building for this future:
- Agents that build other agents, identifying new areas for automation.
- MCP servers evolving with workload performance and planning agents.
- Semantic mapping and insight generation agents that bring industry data models to life.
- Compute, storage, and workload agents that optimize scaling, caching, and schema evolution.
- Performance and operations agents that streamline platform management.
The shift from human-driven analytics to agentic AI will massively increase workloads:
- A basic agentic query can generate 10× the workload of a human query.
- Outcome-based reasoning can drive 25× more queries per minute.
- There will be not just one agent, but potentially ten times more agents than humans.
This requires platforms that can handle extreme concurrency, and mixed workload requirements that could be delivered anywhere; on-premise, edge or in the cloud — where Teradata’s architecture and workload management capabilities excel.
Moving from Experimentation to Impact
The key message is clear: AI success in telecoms is no longer about pilots or isolated use cases. It depends on strong data foundations, integrated architectures, and clear alignment with business outcomes.
As hybrid data cloud and agentic AI adoption accelerates, telcos that invest early in scalable, secure, and interoperable platforms will move beyond pilots—turning AI into a core engine for efficiency, innovation, and long-term competitive advantage.
Get in touch
Get in touch with Laurent Laisney to book a discovery call and consultation session tailored to your organization’s AI and data strategy.