Decision intelligence emerges as telecom’s control layer for AI modernization – News

Market data predicts that by 2027, more than 50% of business decisions will be augmented or automated by AI-driven decision intelligence platforms. That shift reflects a structural transition from insight generation to outcome orchestration.

In the latest episode of theCUBE Research’s AppDevANGLE podcast, Paul Nashawaty, principal analyst at theCUBE Research, spoke with Sreedhar Rao, global chief technology officer for the telecom vertical at Snowflake; Joji Philip, global industry GTM lead for telecom at Snowflake; and Molham Aref, chief executive officer of RelationalAI. Their discussion explored how decision intelligence is emerging as the control layer that transforms telecom data, AI and programmable networks into measurable business outcomes.

“If running your business was like driving a car, business intelligence is the dashboard,” Aref said. “Decision intelligence is the navigation system. It figures out the best way to get to your destination.”

From dashboards to decisions

Telecom operators have invested heavily in business intelligence, observability and AI. Yet BI primarily answers what happened and what is happening. Decision intelligence shifts the focus to what should happen next. In telecom environments, that means moving beyond dashboards that show congestion toward systems that dynamically reroute traffic, allocate spectrum or adjust network slices in real time. It also extends to predictive maintenance, churn prevention, fraud detection, infrastructure planning and energy optimization.

“The decisions that must be automated are the ones that directly control outcomes — SLAs, customer experience, cost, risk,” Philip said. “And that needs to happen at a speed humans simply can’t match.”

This shift is particularly relevant as connectivity evolves into a programmable platform. APIs such as quality-on-demand expose network capabilities directly to developers and enterprise customers. But programmable access without intelligent orchestration risks becoming little more than a slogan.

“Decision intelligence is where programmable networks stop being a slogan,” Philip added.

Closed-loop automation versus auditable intelligence

Automation in telecom is not new. Closed-loop systems have long triggered predefined actions based on thresholds or rule sets. The difference today lies in the complexity of cross-domain tradeoffs.

“The tradeoffs in traditional automation are baked in,” Rao said. “But in a dynamically changing environment, predetermined rules aren’t sufficient.”

Decision intelligence platforms reason across competing objectives, such as performance versus energy efficiency or latency versus cost, and provide auditability into why a decision was made and which alternatives were discarded. That audit trail becomes critical as operators pursue Level 4 and Level 5 autonomous network operations.

In programmable environments, decision intelligence enforces entitlement policies, orchestrates resource allocation across domains and provides proof of SLA fulfillment. If an outcome is not delivered, the system can trace why and trigger corrective action.

Governance, speed and architectural choices

One of the persistent challenges in deploying advanced AI systems has been governance complexity. Implementing decision intelligence has previously required additional technology stacks, duplicated data pipelines and separate security models. 

According to Aref, embedding decision intelligence directly inside Snowflake’s native architecture removes much of that friction.

“No data leaves the security perimeter,” he said. “It’s governed the same way, secured the same way and integrated directly into enterprise workflows.”

This architecture allows business intelligence, decision intelligence and generative AI agents to operate coherently in a single environment. That alignment reduces integration overhead and lowers the operational risk of cross-platform data movement, which is a critical factor for regulated telecom environments.

Rao emphasized that the architectural foundation must begin with streamlined data acceleration, followed by a knowledge plane that abstracts enterprise semantics from raw telemetry. On top of that, decision logic must be exposed as a service rather than embedded inside monolithic applications.

“What you need is decision-as-a-service,” Rao said. “A layer that agents and applications can call into — logical, reasonable, auditable and deterministic.”

Bridging generative AI and measurable value

Frontier large language models continue to advance rapidly. Yet many enterprises struggle to convert generative AI experimentation into measurable operational impact.

Aref described a widening gap between model capability and enterprise value realization. Applying generative AI to business intelligence surfaces insights, but coupling it with decision intelligence enables prescriptive reasoning, graph reasoning and rule-based optimization.

“When you bring BI and decision intelligence together, that’s the unlock,” he said.

In practice, this means reducing operating expenses through automation of repetitive tasks, increasing revenue through optimized offers and churn reduction, improving quality of service and accelerating decision cycles. Generative AI can also assist in building semantic models, predictive features and evaluation frameworks, lowering the historical labor barrier to implementing decision intelligence at scale.

Decision intelligence: The bottom line

Telecom operators are under pressure to deliver programmable connectivity, sustainability gains, autonomous operations and revenue growth. Data alone is no longer sufficient. Even AI-generated insight is not enough.

Decision intelligence is emerging as the control plane that connects data, AI and infrastructure into outcome-driven systems. For operators heading into events such as Mobile World Congress, architectural choices around governance, knowledge abstraction and decision-as-a-service models may determine whether AI remains experimental or becomes operational.

As Aref put it, “The opportunity isn’t just to know what happened. It’s to reason about what should happen next — and act on it.”

Watch the full conversation with Sreedhar Rao, Joji Philip and Molham Aref on theCUBE Research’s AppDevANGLE podcast series to explore how decision intelligence is reshaping telecom AI modernization:

Image: News