AIORG-W002 Working papers

Mapping AI into Production

INSEAD Working Paper 2026/20/STR · SSRN 6513481

working paper abstract

Recorded claims
2
coded from the inspected source
Evidence classes
1
scholarly-metadata
Recorded limitations
4
stated, never hidden

From the wave-0 evidence ledger

Recorded claims

Each claim states what the inspected source says, at the recorded location—bounded by its scope and graded by its confidence. Nothing here is a synthesis across works.

Providing startups with concrete examples of how peer firms reorganized production around AI increased reported AI use cases by 44%, AI-supported tasks by 12%, the probability of acquiring a paying customer by 18%, and revenue by a reported factor of 1.9 during the study window.

confidence: medium-high scholarly-metadata

Scope: 515 high-growth accelerator startups over a short post-treatment window; not mature-firm or long-run evidence.

The intervention reduced reported capital demand by 39.5% while leaving labor demand statistically unchanged, which is inconsistent with treating near-term AI production reconfiguration as mechanically equivalent to headcount substitution in this setting.

confidence: medium-high scholarly-metadata

Scope: Startup accelerator participants; demand intentions/near-term behavior, not economy-wide labor effects.

Boundaries

Limitations & independence

Recorded at coding time, carried with the work forever. A claim without its limits is not evidence.

Recorded limitations

  • 515 high-growth startups in one accelerator context
  • Approximately three-month horizon
  • Revenue distribution is skewed and durability is untested
  • Working paper

Source independence

Academic field experiment in a startup accelerator; intervention and outcomes are researcher-reported.