For more than a decade, connected products have promised to deliver greater value by anticipating failures, accelerating resolution, and improving uptime across growing fleets. OEMs invested accordingly, deploying sensors, building dashboards, modernizing data platforms, and funding digital initiatives at scale.
Yet for many organizations, the returns never matched what was expected.
Recent data reinforces what product, service, and digital leaders already experience firsthand. Gartner reports that only 48% of digital initiatives meet or exceed outcome targets, and PwC found that 92% of operations and supply chain leaders say technology investments have not fully delivered expected results. The issue is rarely a lack of data or tooling. The problem shows up later, when insight still has to be turned into action.
Outcomes, not visibility
Most connected product strategies were justified by outcomes: fewer truck rolls, faster diagnosis, or higher uptime. The path to recurring revenue was built on delivering measurable performance, not just shipping equipment.
In practice, many programs have produced visibility without operational leverage. Alarms fire, but teams do not trust them. Dashboards are refreshed with the latest information, but someone must interpret it. Even when data is technically available, it often requires the same manual review, prioritization, and judgment it was supposed to eliminate.
As fleets scale, this breaks down in predictable ways. False positives create alert fatigue until teams become desensitized. False negatives hide significant issues until failures occur. Trust in connected product signals erodes, and the work of interpretation often shifts to the operator or end customer. The organization ends up with more data, but the same number of decisions need to be made.
Scale exposed the real constraint
The real bottleneck appears after data is captured when signals must be interpreted and decisions made.
Every meaningful action still depends on an expert stitching together telemetry, events, logs, historical behavior, and situational context to decide what is happening, what matters, and what should happen next. That expertise sits with a limited set of reliability engineers, support teams, and technicians. Their capacity does not increase simply because the connected fleet does.
The number of situations that could be acted on grows quickly, while the number of people who can act does not. As a result, many OEMs analyze only a fraction of their deployed base because the manual effort is too high. That creates a ceiling on connected product value that no amount of additional instrumentation can fix.
Workforce dynamics make this even harder to ignore. Deloitte estimates 3.8 million manufacturing jobs will be needed from 2024 to 2033, with 1.9 million potentially unfilled. Scarcity makes expert time more expensive, especially in workflows that still require humans to validate alarms, interpret charts, and decide what to do next.
Data accessibility does not equal action
In response to pressure to get more value from IoT investments, many organizations have modernized their data stack. They migrate platforms, normalize telemetry, build data lakes, and layer on analytics. These investments improve data quality and accessibility, but they do not automate decisions. They do not replace the human judgment that closes the loop.
Even newer AI investments tend to stop short of closing the loop. Machine learning has been primarily applied to anomaly detection, flagging that something is different without determining what it means or what should happen next. Large language models are often layered on top as chat-based copilots, summarizing data or retrieving information only when prompted by an expert. These approaches remain dependent on human interpretation, leaving the core constraint unchanged: expert judgment still sits in the critical path for every decision.
Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data. Even as data maturity improves, the operational gap remains. Wavestone’s 2024 executive survey shows 48.1% of organizations now report having created a data-driven organization, up from 23.9% in 2023. That progress is real, but being data-driven does not automatically mean decisions are being made faster, more consistently, or at scale.
Operators are not waiting for OEMs to solve this problem. Many are adopting broad operational platforms and building their own action layers on top of OEM data feeds. When an OEM’s connected product experience stops at alarms and dashboards, the OEM risks becoming a commodity signal provider rather than a differentiated outcomes partner.
In practice, the organizations that pull ahead are not the ones with the most dashboards, the newest platforms, or the most polished AI summaries. They are the ones that reliably close the loop between sensing and acting, across the full fleet, without requiring a constant human review cycle.
From signals to outcomes
What is becoming clear is that the ceiling on connected product value is no longer set by sensing, connectivity, or data availability. It is set by the ability to apply expert judgment at scale. As fleets grow and operating environments become more complex, systems designed primarily to surface information struggle to deliver consistent outcomes because decision-making remains constrained by human capacity.
This gap is now shaping an emerging category. Rather than stopping at detection, visualization, or summarization, these systems focus on replicating expert reasoning in software. AI agents are engineered to retrieve relevant data, apply situational context, and determine appropriate actions for specific operational scenarios. When rigorously designed and evaluated, they can be deployed directly into production workflows and execute high-value actions without requiring a human in the loop.
Closing the connected product expectation gap depends on this transition: from systems that inform to systems that act, and from insight as an intermediate step to outcomes as the objective.
Kurt Neuens is Vice President of Product & Strategy at Mesh Systems.
Learn more about MeshInsights, our solution to applying expert judgment at scale across connected products: