Building resilient innovation systems in an age of continuous disruption

Last summer, I had the opportunity to speak about uncertainty becoming a constant feature of the RD+I operating environment.

At the time, the point was not that disruption was increasing, but that many organizations were still structured as if disruption were episodic. Innovation was something to ramp up during moments of change and optimize during periods of stability. But, that model was already starting to break down.

The mismatch between how organizations were designed and how they actually needed to be operating was becoming increasingly visible. A year later, that gap has only widened.

The pace of Artificial Intelligence (AI) advancement, geopolitical fragmentation, faster competitive cycles, evolving talent  dynamics, and the pressure these forces are placing on existing operating models have reinforced what was already clear. Uncertainty is not something organizations can move through. It is something they operate within. And yet, in many cases, the underlying systems have not kept up.

Reach out to discuss this topic in depth.

From innovation initiatives to innovation systems

What I discussed then still holds, but it becomes clearer when grounded in how organizations actually operate. Many still approach innovation as a set of initiatives, while the underlying ways they make decisions, allocate resources, and execute work remain unchanged.

A resilient innovation system is not simply one that performs well under pressure. It is one that is structurally designed to adapt. In practical terms, this means being able to sense change early, reconfigure resources and priorities without excessive friction, and execute decisions at scale without introducing new complexity.

Many organizations continue to invest in innovation capabilities without addressing the underlying ways of working that limit their impact. As a result, they generate ideas but struggle to adapt how those ideas are prioritized, resourced, and executed as conditions change.

Organizations that are making meaningful progress are approaching this differently. They are not just adding innovation initiatives. They are redesigning how the system itself operates so that adaptation becomes continuous rather than episodic.

Rethinking planning and the innovation operating model

This shift becomes visible when looking at how organizations plan and how innovation work is structured.

Traditional planning approaches were built on the premise that, while the future was uncertain, it would remain stable enough to support decisions over defined time horizons. That premise is becoming harder to sustain as the pace and variability of change increases. This does not eliminate the need for stability in execution, but it does change where and how flexibility needs to be built into the system.

One organization I profiled in my discussion was beginning to experiment with more dynamic approaches, using simulation environments to test hypotheses, explore potential outcomes, and better understand how decisions play out within specific parts of the Research & Development (R&D) process. These efforts are still early, but they point to a different way of operating: one where planning becomes less about committing to a fixed path and more about building a system that can adjust quickly as new information becomes available.

In most organizations, however, planning remains largely unchanged. New tools may be introduced, but the underlying cadence and decision structures persist. As a result, teams often have more data available to them, without a corresponding shift in how quickly or effectively they can act on it.

A similar shift is needed in how innovation work itself is organized. In the talk, I described ecosystems as adaptive, decentralized, co-creative, and co-evolving. In this context, ecosystem is not just a reference to external partnerships, but a contrast to more traditional, hierarchical models where work is structured in fixed functions, decisions flow through layers, and coordination depends on centralized control.

An ecosystem-oriented approach rethinks this structure. Work is organized more flexibly, decision-making is pushed closer to where information sits, and teams form around specific problems rather than remaining fixed within functional boundaries.

In one of the cases discussed, this was applied directly to the R&D operating model. Work was broken into smaller, self-directed units, with accountability and decision-making embedded within those units rather than managed centrally. This allowed teams to reconfigure more quickly as priorities shifted, particularly in environments where work spans multiple disciplines and time horizons.

This kind of model is not widely adopted, and it is not easy to implement. However, it illustrates what a more adaptive approach to organizing R&D and innovation work can look like in practice.

Resilience, in this context, depends not only on how organizations plan, but on how work itself is structured. Systems that rely on fixed hierarchies and centralized coordination tend to adapt more slowly, while more modular and distributed models can adjust more readily as conditions change.

From AI support to Systems of Execution (SoE)

The most significant shift over the past year, however, has been in the development of more advanced AI capabilities, particularly in the form of agentic systems.

When I first gave this talk, AI was already influencing how decisions were made, primarily as a support tool. Since then, we have started to see the emergence of systems that can go further by coordinating workflows, interacting with tools, and executing parts of multi-step processes within R&D environments.

These developments are still evolving, but they point to a different way of thinking about how innovation work can be carried out. Rather than focusing only on augmenting individual decisions, the emphasis shifts toward enabling more continuous, coordinated execution across tasks.

This is where the idea of scaling intelligence becomes more concrete: human and AI capabilities no longer operating in isolation but increasingly integrated into workflows that can manage complexity over longer time horizons.

In this context, the challenge is not only how to use AI effectively, but how to structure work so that these capabilities can be applied in a consistent and scalable way. As that happens, the ability to accelerate innovation depends not just on individual expertise, but on how well intelligence can be orchestrated across the system.

Measuring and designing for resilience

As innovation systems become more dynamic and more distributed, the way organizations measure success also needs to evolve.

Most metrics in use today remain lagging indicators. They provide useful information about outcomes, but limited insight into whether the organization is becoming more capable of adapting to disruption. In contrast, leading indicators of resilience focus on the conditions that enable adaptation.

In the discussion, one example highlighted how this can be approached more systematically, using dimensions such as willingness to learn, adaptability, collaboration, and shared vision to assess how teams respond under changing conditions. Rather than measuring performance after the fact, this kind of approach provides earlier signals of where the system is strengthening and where it may be under strain.

This does not replace traditional metrics; rather, it complements them by shifting some of the focus toward building the capabilities and conditions that enable those outcomes.

Designing for continuous adaptation

Taken together, the shifts outlined last year point to a broader change in how innovation needs to be structured. It is no longer just about generating ideas or accelerating individual initiatives. It is about reshaping how organizations plan under uncertainty, organize work, apply intelligence, and measure progress.

These elements are closely connected. More dynamic planning depends on more flexible ways of organizing work, while advances in AI are changing how execution can be coordinated and scaled. At the same time, new approaches to measurement are reshaping how organizations understand their ability to adapt.

Viewed individually, each of these changes can seem incremental. In combination, they begin to alter how innovation operates.

Resilient innovation systems are not built through a single capability or tool. They emerge from how these elements are designed to work together. That, ultimately, is what determines whether an organization can continue to adapt as conditions change.

If you enjoyed this blog, check out, Our convictions for R&D and Innovation in 2026  – Everest Group Research Portal, which delves deeper into another topic relating to R&D.

If you’d like to continue this discussion, please contact Jillian Walker ([email protected]).