Innovation Economics · Heatmap
Sen's capability matrix.
Amartya Sen reframed poverty and development as a question of substantive freedoms — what people are actually able to do and to be — rather than income or utility. The capability frame asks a different question of any AI deployment: what does this expand the capability to do, and for whom? An AI tool that automates a banking process may expand the capability of the bank's customers (faster service, thin-file access) or contract it (foreclosed services for those the model excludes). Same tool, different outcomes by design and deployment.
| Deployment | Health | Education | Voice | Livelihood | Security | Inclusion |
|---|---|---|---|---|---|---|
| Rural diagnostic AI | + | · | · | + | + | + |
| Indigenous-language agents | · | + | + | · | · | + |
| Banking automation | · | · | − | · | − | − |
| Thin-file credit scoring | · | · | · | + | · | + |
| AI hiring screening | · | · | − | − | · | − |
| Township fintech AI | · | · | · | + | · | + |
| School-tier tutor AI | · | + | · | · | · | + |
| Predictive policing | · | · | − | · | − | − |
| Smallholder yield AI | + | + | · | + | + | + |
| Generative content tooling | · | · | + | · | · | · |
The thesis
The capability lens highlights that innovation can reduce income poverty while leaving capability poverty unchanged — or even expanding it, if the technology degrades the capabilities of marginalised groups. Indigenous-language coverage in AI agents is a capability question, not a feature question. Access to AI-assisted healthcare in rural areas is a capability question, not an infrastructure question. For an MEA-headquartered enterprise, Sen's frame is the most useful single lens for distinguishing AI deployments that genuinely serve from those that merely sell.