Rwanda RapidSMS
Map a referral from the community health worker to the facility. Compare whether a message was received with whether the family received care.
From challenge to capability
A working approach to developing capability: a definition, a seven-step cycle and questions that make each step useful. Refinement continues throughout rather than waiting for a final stage.
Learning engineering connects human flourishing, learning sciences and systems engineering through evidence-informed iteration. In LENS, this means examining what people can do together with the tools, authority and opportunities that enable action. The cycle makes those assumptions and evidence needs explicit; completing its steps alone does not demonstrate learning or impact.
Bring human flourishing, learning sciences and systems engineering to the same challenge. Understand the work, change the conditions for action, and use evidence to revise decisions. The pipeline offers seven steps with 8. Refine throughout.
A working definition
The capacity of a person, team, or human-system arrangement to accomplish valued work under stated conditions, with usable opportunities to act.
We examine output quality, capability, learning, and impact as distinct claims. Each needs evidence that fits the work, its conditions, and the decision being made. A good output alone does not establish that someone learned, can perform independently, or can act in another setting.
William Gray-Roncal & James Diamond · Working definition for community discussion, developed in Capability Matters.
Three foundations, one shared purpose
Develop capability that serves people, rests on credible learning evidence, and works in the conditions where it matters.
Our synthesis connects these fields. OECD’s Education for Human Flourishing (2025) informs the human-purpose component. Its Learning Compass also connects agency with individual and collective well-being.
Start with a challenge. Follow Understand → Map → Design → Build → Instrument → Experiment → Evaluate, then return directly to Understand. Refine any step when evidence changes an assumption; the numbered sequence is a guide to decisions, not a rule to delay feedback.
On a narrow screen, scroll the diagram horizontally. The numbered explanations below provide the same sequence in text.
Frame the capability gap before choosing a solution. A sponsor’s target can reward activity while leaving the actual problem unchanged.
Decision cue: Distinguish what people need to do from what is easy to count.
Ask: Whose task, access conditions or definition of success have you not checked?
Map capabilities and conditions. People, tools, authority, incentives and support jointly shape what is possible. A skills diagnosis alone can miss a system constraint.
Decision cue: Connect human and system requirements, dependencies and competing explanations.
Ask: What observation would distinguish a knowledge gap from an access, workload or tool problem?
Compare interventions and their tradeoffs. A useful intervention acts on a plausible mechanism within your authority; adding features does not resolve an invalid assumption.
Decision cue: Compare a simpler alternative, name who keeps control and identify a stop condition.
Ask: What does this choice exclude, and who carries the burden if it fails?
Make the smallest useful candidate real. Construction reveals integration, staffing and recovery requirements that a plan can hide.
Decision cue: Exercise a representative task and its failure path before expanding scope.
Ask: Which dependency, fallback or correction path must actually work?
Choose evidence that can change a decision. A convenient metric can improve while the capability you care about stays unchanged.
Decision cue: Choose the outcome, comparison, missing-data checks and decision rule before reading results.
Ask: Could these measures reveal that your preferred approach failed, or exclude an important group?
Test the instrumented intervention under representative conditions. A bounded experiment can reveal how access, staffing and support affect the result before wider use.
Decision cue: Specify the comparison, permissions, support and stopping rule for the test.
Ask: What evidence would justify changing the intervention or moving toward deployment?
Interpret evidence within its limits. A favorable trend can reflect chance, selection, exposure or another change; a null can reflect an ineffective intervention or limited evidence.
Decision cue: Separate observed change from causal explanation and unmeasured outcomes.
Ask: What claim is supported, for whom, over what period, and what remains unknown?
These cases show how different kinds of work call for different evidence.
Map a referral from the community health worker to the facility. Compare whether a message was received with whether the family received care.
Use local inquiry to design an advising tool. In a pilot, observe whether advisers can act on a flag and whether students can challenge a mistake.
Compare simulator assessment with later performance during supervised work at sea. Use the difference to refine practice and qualification.
Worked narratives and a fictional simulation offer different ways to examine these decisions. Each has its own instructions, evidence limits and controls.
Study a narrative, find the turning point, or predict before revealing the explanation. Compare how different system constraints change the decision.
Make decisions in a fictional scenario, inspect outcomes, and compare revised branches. Its probabilities are teaching assumptions, not forecasts.