Jevons Paradox
person concept tool org talk claim — click a node to jump to its page; hover an arrow for the relation
An economic principle holding that efficiency improvements in resource consumption tend to increase total demand for that resource rather than reduce it, because lower costs expand the scope of use.
Application to AI Agents and Data Access
In the context of AI agent engineering, Jevons Paradox is invoked to explain a counterintuitive outcome when agents reduce the friction involved in cross-team workflows. As discussed in Inside YC's AI Playbook, when agents remove the back-and-forth overhead between teams required to get something done, the result is not simply that the same work gets done faster — instead, users dramatically increase both the quantity and complexity of the questions they dare to ask. The efficiency gain expands total demand rather than merely satisfying existing demand more cheaply.
This dynamic is particularly relevant to data query workflows: once an AI agent lowers the cost of asking a question of another team (or a data system), individuals who previously self-censored requests as "not worth the effort" begin making those requests, and they also escalate to more ambitious, compound questions. The net effect is that the agent generates more inter-team coordination activity in aggregate, not less — a classic Jevons effect.
Implications for Agent Design
This framing suggests that evaluating agents purely on latency or cost reduction understates their impact. Builders and organizations deploying agents for internal tooling or data access should anticipate demand expansion as a primary outcome, with downstream consequences for infrastructure scaling, evals coverage, and governance — since the tail of newly-asked complex questions may surface edge cases that simpler, lower-volume workflows never encountered.