Curated data can give communications service providers (CSPs) a more useful basis for AI than raw packet volume alone. By correlating subscriber sessions with location, timing, network domain and performance indicators, a CSP can expose information that is easier to query and act on. Rick Fulwiler, senior director of product management at NETSCOUT, proposes three applications: predictive maintenance, capacity planning and third-party services. They are industry examples, not independently demonstrated results.
What curated data means for a CSP
Network equipment produces vast amounts of traffic and operational information, but volume is not the same as context. An AI system looking at isolated packets may not know which subscriber session they belong to, where and when it occurred, whether it crossed the radio access network, core or transport domain, or how its performance related to service indicators.
As an Amazon Associate I earn from qualifying purchases.
In Fulwiler’s August 24, 2026 article in The AI Journal, curation means correlating packet-level traffic with information such as session lifecycle, location, timing, session flow, network domain and key performance indicators. The result is a structured view of network state that an AI system can query. This is contextual enrichment, not simply filtering or compressing data.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteThe distinction matters because an illustrative estimate in the article puts raw user-plane traffic for a 5G Standalone network with 10 million subscribers at about nine petabytes per day, near one million transactions per second. That is Fulwiler’s example, not an independently verified benchmark. It illustrates the scale problem: feeding everything into an AI workflow may be less useful than making relevant relationships and state available.
#1 Best Overall
How curated data could support predictive maintenance
Correlated session and network information could help an operations team look for changes that precede a familiar alarm. Fulwiler’s examples include abnormal state transitions, registration failures, traffic degradation and drift in quality of service. Considered together, such signals might indicate an emerging service or capacity issue rather than a collection of unrelated events.
In the proposed workflow, an AI agent identifies a risk from that contextualized information and recommends an intervention. One example is scaling capacity for a 5G Standalone user-plane function (UPF) ahead of expected demand. The article presents this as a possible operational use, not evidence that a deployed system predicted failures or prevented outages.
For a CSP evaluating the idea, the important test is whether an alert or recommendation arrives early enough to change an operational decision. Useful evaluation would connect the agent’s signals to incident records, false alarms, missed events, response time and the outcome of any recommended action. The article does not report those measurements.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →How curated data could improve capacity planning
Capacity planning requires more than a current utilization snapshot. A CSP could combine historical signaling and traffic-volume patterns by market, month and event type to identify where demand tends to rise and which network resources may become constrained. That contextual view could inform capacity plans and capital spending before a predictable demand peak.
This remains a proposed use case, not a reported forecasting result. Fulwiler’s article supplies no forecast-accuracy figures, cost savings or deployment case study. A CSP considering the approach would need to compare forecasts with actual demand and track whether they improved decisions across markets and event periods.
How third-party applications could query network context
Fulwiler describes a hypothetical mobile-banking scenario involving international transfers. The application needs context about a subscriber’s location and session integrity. The article says direct Home Location Register (HLR) queries and ingesting raw call-detail records or packet data were inadequate for this imagined workload; it proposes exposing a narrower, curated view instead.
Rank #4
- NOTE: Compatible Only with FALA IOT monitors and Y-splitters, exclusive FALA IOT ecosystem integration
- Length: 16.4 feet ( 5 meters), flat cable type. Custom lengths and special features available
- Ultra-wide Monitoring Range: The digital temperature probes & sensors measures extreme temps from -40°F to 248°F with lab-accurate ±0.6°F precision
- Anti-slip Design: Stays securely in place for uninterrupted monitoring. It'll work well through vibrations, handling, and harsh work environments
- High Quality: The premium stainless steel probe is waterproof & rust-proof for reliable performance in farming, cold chain, refrigerator, labs, drying box, constant box and industrial use, etc.
The proposed architecture collects and curates information at its source, including through deep packet inspection (DPI), then makes the structured view available to an AIOps platform, AI agent or third-party application. The article presents Model Context Protocol (MCP) as a possible standardized gateway between the curation layer and agents or applications. MCP is an integration proposal here, not proof of interoperability, lower cost or improved reliability.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
In the banking example, the author reports “second-level resolution” and a better user experience. No bank, study design, baseline or independent results are provided, so those outcomes should be understood as claims within the example rather than established performance. For real integrations, CSPs would also need to define which fields an application may access, how identity and authorization are handled, how fresh the information must be, and how access is audited.
Best Value
What CSPs should assess before adopting a curation layer
The article does not compare products or document an implementation. CSPs can still use its proposal to frame an evaluation around the operational requirements that determine whether contextual data is useful:
- Scope and provenance: Specify which packet, session and KPI fields are included, where they originate, and how the data is validated.
- Correlation: Check whether the system can reliably link subscriber sessions across relevant network domains and lifecycle states.
- Freshness and latency: Set acceptable delays for each workflow; a capacity forecast and a real-time service check may not need the same update rate.
- Privacy and access: Define data minimization, authorization, retention and audit controls, especially when third parties can query network context.
- Interoperability: Test how the curated data is exposed to existing AIOps systems, agents and applications, including whether MCP fits the required interfaces.
- Operational integration: Determine how recommendations reach network teams, what actions require approval, and how automated changes can be rolled back.
- Measured outcomes: Set baselines and success criteria for incident detection, forecast accuracy, response time and service impact before deployment.
Other possibilities—and what the examples establish
Fulwiler also names digital twins, energy-efficient operations, dynamic slice management, closed-loop automation, security and fraud detection as possible extensions. These are prospective applications, not use cases evaluated with results in the article.
The central proposition is that AI systems may reason more effectively when network information is organized around sessions, context and service state rather than supplied as unprocessed traffic. The three examples show how that proposition might shape operational and third-party workflows; they do not establish measured gains or prove that a particular architecture will deliver them.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
