Persona Migration and Expectation Recalibration in Generative AI Adoption: A Longitudinal Study at a State Department of Transportation
Citation. Shoghli, O., Banani Ardecani, F., & Mohamadi Hezaveh, A. (2026). Persona Migration and Expectation Recalibration in Generative AI Adoption: A Longitudinal Study at a State Department of Transportation. arXiv. https://arxiv.org/abs/2607.13798
Summary. This preprint follows employees through an eight-week Microsoft 365 Copilot pilot in a state transportation agency, measuring perceived usefulness, ease of use, behavioral intention, and trust before and after use. It finds that aggregate usefulness declined after hands-on experience while individual adoption personas showed substantial migration in both directions.
Key claims
- In a matched two-wave sample of 124 employees, perceived usefulness declined significantly after the pilot, while aggregate change in ease of use, behavioral intention, and trust was small and non-significant.
- Forty percent of baseline Skeptics moved to a Cautiously Positive persona, while 68% of baseline Champions moved to less enthusiastic personas, so aggregate averages obscured substantial individual change.
- The authors argue that AI adoption support should be dynamic and persona-specific, using task-relevant examples, verification routines, and trust calibration rather than treating pre-use enthusiasm as a stable forecast.
Evidence & limitations
- This is a single-agency, eight-week preprint study without a control group, so changes should not be interpreted as causal effects of Copilot alone or generalized uncritically across organizations.
- Persona clusters and open-ended keyword mapping are exploratory; the work has not yet undergone peer review.