Development 2026: It’s Productivity, Stupid
The question for the development sector today is not whether aid should exist, whether technology will transform society, or when public and political priorities will shift again. The question is colder: how to deliver outcomes in a world of abundant information, distributed expertise, tighter funding and shorter political patience.
This is a practical problem, not a strategic one. Development institutions were designed for a world that no longer exists: scarce information, concentrated expertise and generally expanding resources. They now operate fully in public view, judged not by intention, mandate or reputation, but by their ability to demonstrate value.
Most development organisations have evolved through the slow accumulation of processes, reviews, reporting requirements and layers of management. Each may have been defensible. Together, they now convert too many resources into administration before they convert them into outcomes. Doing better is productivity, not politics.
Productivity does not mean working harder, cutting standards or weakening safeguards. It means translating resources into impact more effectively. It applies to process, structure, systems, cost, staffing, procurement, project preparation and knowledge management. Every layer of activity should justify its contribution to outcomes, not just control. Technology, including artificial intelligence, sits at the core of this redesign.
Development’s greatest assets are the human experience and knowledge accumulated through decades of projects, failures, adaptations and successes. The sector’s weakness lies in how poorly it uses these assets. Knowledge remains fragmented, hard to retrieve, poorly reused and often repeated. Staff spend too much time navigating systems rather than applying judgement. Expert input often serves as a process step rather than an action driver.
With visible outputs and immediacy now part of the operating context, development organisations should start with them and work backwards. What decision must be made? What expertise does it require? What controls genuinely protect integrity? What can be simplified, automated or removed? Strong institutions remain essential, but strength must now be combined with agility.
In this context, artificial intelligence matters because it serves the task, not because it is fashionable or replaces human judgement. It reduces administrative effort, improves institutional memory, accelerates analysis and moves scarce expertise towards higher-value work.
Development does not need fewer institutions. It needs lighter, more capable institutions that demonstrate value in both the short and long term. A focus on productivity strengthens development’s ability to show value now and over time, regardless of changes in public or political priorities.
