Evidence Record

Evidence summary

DORA reports that heavier AI use correlates with better self-reported flow, satisfaction, and productivity while a 25% adoption increase associates with 1.5% lower delivery throughput and 7.2% lower delivery stability.

Exact claim supported or contradicted

Individual experience and system-level delivery performance can move in opposite directions.

Source provenance

Primary organizational research report using large-scale survey/interview methods; this record relies on the publisher's public summary.

Relevant excerpt or data

Reported associations: +25% adoption, -1.5% throughput, -7.2% stability.

Interpretation

Local code generation gains may shift costs into batch size, review, integration, or operations.

Limitations

Observational associations do not establish causality; constructs and sample composition require full-report review; self-report and common-method bias.

Counterevidence

Controlled task evidence can show direct speed gains, but does not measure the organizational outcome.

Reproduction or verification notes

Test using repository telemetry with adoption timing, batch-size mediation, quality outcomes, and matched controls.