Stop Building Better Tools—Start Fixing the Spaces Between Them

We automated tasks, not trust. That’s why nurses sit idle in one hospital while another pays agencies double.

Picture a hospital network where nurses at Site A work overtime while Site B desperately calls expensive agencies because a nurse phoned in sick. The technology exists to solve this in minutes. The nurses exist within the network. Yet it doesn’t happen. Why?

A nurse manager at a metropolitan hospital, experiences this daily. The hospital might have incredible systems for tracking patient care, managing medications, even predicting which patients might deteriorate, But when three nurses call in sick on the same shift, the only option is paying agencies double the normal rate — even though there are qualified nurses fifteen minutes away at a sister hospital who’d love the overtime.

This scenario plays out thousands of times daily across healthcare systems worldwide. It’s not a story about insufficient technology or lazy administrators. It’s a window into the real productivity crisis of our time: we’ve become brilliant at optimising within organisations while becoming terrible at coordinating between them.

TL;DR Tech works, but our biggest productivity leaks live between organisations. Fixing them requires governance, not gadgets.

The Coordination Paradox

We have the richest productivity toolkit in history. Large Language Models can automate bureaucratic tasks that once consumed entire afternoons. AI can diagnose diseases, predict equipment failures, and streamline workflows. Voice assistants help healthcare workers deliver better care with improved staffing ratios. Yet despite these impressive capabilities, overall productivity continues to stagnate.

Recent studies reveal a striking paradox. At the task level, LLMs can deliver large productivity gains: randomised trials report improvements ranging from 20% to 80% depending on the activity.⁠[1] Yet when these results are averaged across a worker’s portfolio of tasks, the effect shrinks to a single-digit improvement in overall performance.⁠[2] And at the firm level, researchers find little evidence so far of measurable gains.[⁠3] The technology is working exactly as promised; what these results underscore is that the true bottleneck is no longer individual tasks but the broader systems in which they are embedded.

How We Got Here: The Great Complexity Migration

To understand why coordination has become our Achilles’ heel, we need to look at how business has evolved over the past thirty years. Companies have systematically moved complexity from inside their walls to outside them, transforming themselves from integrated operations into lean coordination hubs.

The poster child for this transformation is category management in retail. Instead of managing product categories internally, retailers shifted this responsibility to suppliers, who now optimise inventory, pricing, and promotions collaboratively. The result? Roughly 30% gains per affected product line. What once required large internal teams became someone else’s specialised expertise.

This pattern repeated across industries. Even capital-heavy sectors like Australia’s superannuation industry — which processes around $5 billion weekly — now operate as networks of small to medium businesses, outsourcing complexity to external providers.

The strategy worked brilliantly for individual organisations. Companies became more agile, focused, and profitable. But it had an unintended consequence: the biggest opportunities for productivity improvement migrated to the spaces between organisations, where coordination mechanisms remain primitive.

Healthcare’s Hidden Constraints

Nowhere is this coordination crisis more visible than in healthcare, where inefficiencies that seem internal to hospitals are actually symptoms of ecosystem-level constraints.

Take the agency staffing crisis. At first glance, this appears to be poor workforce planning by individual hospitals. Dig deeper, and you find a system trapped by artificial boundaries. Hospitals operate as discrete units, each managing its own staffing pool, budgets, and rosters. When demand surges unexpectedly, agencies become the only release valve — despite qualified staff existing elsewhere in the network.

The constraint isn’t regulatory or technical — it’s institutional. Distance matters, but not in the way we typically think. The real distance isn’t the twenty-minute drive between hospitals; it’s the organisational distance between separate budget lines, different IT systems, and competing managers protecting their resources.

Consider bed shortages, another visible symbol of system strain. These are often less about absolute scarcity and more about informational gaps. Patients occupy the wrong beds, facilities remain unaware of nearby capacity, and discharge processes drag due to siloed systems. A network-level view would reveal that beds exist — they’re just not visible when decisions are made.

The Coordination Solution

What would ecosystem-level coordination look like in practice?

Start with dynamic labor allocation. Instead of each hospital managing its own staffing pool, imagine a network-level entity that coordinates labor flows. Nurses could indicate their availability for overtime in their profiles. When someone calls in sick, the system automatically offers the shift to qualified staff across the network, bypassing costly agency markups entirely.

This isn’t fantasy — it’s engineering. The University of California demonstrated this principle with procurement in 2014, implementing system-wide coordination that delivered over $100 million in annual savings. They created what was essentially an “internal Amazon” for university purchasing, transforming fragmented buying into coordinated demand.

The technical barriers are minimal. The coordination barriers are everything. Success requires shared governance, interoperable systems, and incentive structures aligned to collective outcomes rather than individual organisational goals.

Beyond Healthcare: The Broader Pattern

This coordination crisis extends far beyond healthcare. In retail, fragmented returns processing creates massive inefficiencies in reverse logistics. In franchising, pricing optimisation happens at individual locations despite network-wide data availability. In professional services, expertise sits unused in one office while clients pay premium rates for the same knowledge elsewhere in the network.

The pattern is consistent: complexity has migrated to ecosystem boundaries, but our coordination mechanisms remain trapped in an era when everything happened within single organisations.

The Limits of the Internal Focus

This doesn’t mean internal optimisation is worthless. Process improvements, automation, and efficiency gains within organisations still matter. But they’re increasingly subject to diminishing returns. The big opportunities — the ones that can meaningfully move the productivity dial — now live in the spaces between organisations.

Some might argue that regulatory constraints make ecosystem coordination impossible. While regulations do create genuine barriers, many apparent “regulatory constraints” are actually institutional choices disguised as necessities. The difference between “we can’t do this because of regulations” and “we haven’t figured out how to do this within regulations” is often the difference between stagnation and breakthrough.

Others point to the genuine complexity and risk of coordination across organisations. They’re right : it’s harder than internal optimisation. But that’s exactly why it represents the next frontier. Easy coordination problems were solved decades ago. What remains requires collaborative governance models, shared risk frameworks, and new forms of organisational relationship.

Questions for Your Industry

The next time you encounter a “productivity problem,” ask yourself: What invisible ecosystem constraint is causing this? Look beyond your organisation’s walls. Where do artificial boundaries force inefficiency? What coordination failures masquerade as resource scarcity?

Consider your own industry:

• Where does expertise exist in your ecosystem but remain trapped by organisational boundaries?
• What information could dramatically improve decision-making if it flowed more freely between organisations?
• Which “resource shortages” are actually coordination failures?
• How might collaborative governance models address systemic inefficiencies?

The Next Productivity Revolution

Real productivity won’t come from better tools within organisations : we have those already. It will come from reimagining the spaces between them. This means building infrastructure for coordination: shared governance structures, interoperable systems, and incentive models that reward collective outcomes over individual optimisation.

The hospital network where nurses work overtime while agencies profit from artificial scarcity isn’t broken because technology failed. It’s broken because we’ve optimised the parts while neglecting the whole. The next productivity revolution lies in rewiring those connections.

The question isn’t whether your organisation can become more efficient. It’s whether your ecosystem can become more intelligent. That’s where the real leverage lies, waiting for leaders brave enough to look beyond their own walls and smart enough to know that the biggest opportunities often hide in the spaces between.

References

1. Noy, S., & Zhang, W. (2023). Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence.NBER Working Paper №31161. https://shakkednoy.com/Noy%20Zhang%20NBER%20SI.pdf. Showed a ~40% time savings in professional writing tasks, with quality improvements.

2. Brynjolfsson, E., Li, D., & Raymond, L. (2023). Generative AI at Work. NBER Working Paper №31161. https://www.nber.org/papers/w31161. Reported average worker-level productivity increases of ~14%, but much smaller when spread across diverse tasks — often cited as ~3% net gain in broad performance.

3. Waber, B., & Fast, N. J. (2024). Is GenAI’s Impact on Productivity Overblown? Harvard Business Review. https://hbr.org/2024/01/is-genais-impact-on-productivity-overblown. Argues that firm-level productivity effects remain difficult to detect, and warns against extrapolating task-level results to systemic outcomes.