Serge Lucio is the VP and GM of Agile Operations Division, Broadcom Inc.
We’re roughly halfway through 2026, and the enterprise outlook remains incredibly optimistic. Over the past few years, we’ve all experienced the rapid, growing adoption of AI. We’ve seen firsthand how it’s challenging the status quo and fundamentally reshaping how large organizations operate.
However, as businesses rush to capitalize on this transformation, a clear dividing line is emerging. Many organizations are taking an overly simplistic approach to the AI era by essentially slapping a chatbot on top of an existing user interface. While this might provide some incremental, surface-level value, it’s not enough to realize the full, transformative benefits that AI can deliver.
Truly unlocking the potential of AI is about enabling “agentic” models: deploying intelligent AI agents exactly where they sit in the workflow, armed with the right domain-specific intelligence. More importantly, it requires the right data. Simply put, AI without the right data doesn’t work.
To navigate this transformation effectively, enterprise leaders must pivot away from superficial AI wrappers and focus on two core imperatives:
1. We must build systems that ensure the completeness, consistency and accuracy of our underlying data.
2. We must deploy the right kind of agentic models that allow users to ask truly meaningful, complex business questions.
When we look at the modern enterprise, this deep focus on data integrity and intelligent agents must extend across three critical operational pillars.
1. Strategic Portfolio Management: From Silos To Synergy
A primary push for many organizations today is the seamless alignment of high-level strategy with day-to-day execution. However, the data required to achieve this alignment is notoriously siloed.
Historically, enterprise planning tools have sat at the intersection of the business, tasked with harvesting and aggregating fragmented data. Aggregation alone isn’t enough, though. The real power of a modern portfolio management strategy lies in ensuring that data is thoroughly vetted, complete, normalized and accurate.
Once this reliable data foundation is established, organizations can introduce next-generation AI agents. With vetted data, leaders can ask these agents highly critical, nuanced questions about capacity constraints, shifting budgets and project timelines. Crucially, the future of enterprise AI lies in interoperability. By exposing these capabilities through standard protocols, organizations building their own custom agentic systems can seamlessly integrate their internal agents directly with broader enterprise platforms.
2. Network Observability: Taming Architectural Complexity
In the realm of network and systems observability, the primary challenge is sheer complexity. To get a clear, actionable picture of enterprise health, operations teams must integrate fault, performance, flow, user experience and configuration data.
The difficulty is that data spanning different generations of technology comes in vastly different forms—from legacy SNMP to modern gNMI telemetry. Frequently, this data overlaps, making it difficult to find a single source of truth. To make AI work here, organizations must normalize all of this varied telemetry into one coherent, unified system that spans from Layer 2 up to Layer 7.
Once operational data is properly normalized, the application of AI becomes transformative. It enables highly effective, automated root-cause analysis and intelligent triage at scale, moving IT teams from reactive troubleshooting to proactive optimization.
3. Enterprise Automation: Establishing A GenAI Control Plane
For years, large organizations have been layering different automation technologies on top of one another. An enterprise might run legacy scheduling alongside traditional workload automation while also heavily using modern data pipeline automations.
These “islands of automation” create severe governance challenges. Managing service-level agreements (SLAs), controlling costs and ensuring that data is delivered exactly when needed becomes incredibly complex. For years, the goal was simply to visualize this entire automation ecosystem through a single pane of glass.
Today, that goal has evolved. While understanding costs and SLA management remains vital, the GenAI era has made data provenance the ultimate priority. This involves strictly managing personally identifiable information (PII), ensuring regulatory compliance and surviving rigorous audits.
Moving forward, the enterprise requires a true control plane for the GenAI era—one equipped with deep data provenance. This provides absolute control over where your data goes, ensures strict compliance and tightly manages exactly what information is shared with autonomous AI agents.
The Year Of The Data Foundation
If there’s one defining lesson for enterprise technology in 2026, it’s that reliable data is the engine that makes AI work for the enterprise.
By stepping beyond the chatbot and focusing deeply on data completeness, normalization and provenance, organizations can build the foundation required to power true agentic AI. Armed with these capabilities, leaders are perfectly positioned to meet their business goals in entirely new ways, ultimately empowering IT teams to become strategic orchestrators rather than reactive troubleshooters.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
