AI Training for Micro-Entrepreneurs

How to cite

Ronconi, Juan Pedro; Balmaceda, Marcos. 2025. "AI Training for Micro-Entrepreneurs." AEA RCT Registry. AEARCTR-0017023. https://doi.org/10.1257/rct.17023-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Juan Pedro Ronconi* , Marcos Balmaceda

* Primary investigator

Registry ID
AEARCTR-0017023
DOI
https://doi.org/10.1257/rct.17023-1.0
First registered
Last updated
Status
In development
Trial dates
2025-08-11 – 2026-10-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/17023

Abstract

The objective of the experiment is to test whether a short online course that introduces GenAI for microentrepreneurs and teaches key business practices for small firms—marketing, inventory management, record-keeping, and financial planning—through examples applying AI prompting skills, leads to greater AI adoption and better firm performance. We collaborate with a Chilean platform that offers free online training to microentrepreneurs to design and implement the course.

Intervention

The intervention consists of offering the course to half of the respondents of the baseline survey. The control group will not receive anything while the trial lasts (pure control). They will receive access to the course after the endline survey is collected (6 months after the intervention).

Experimental design

We obtained baseline data on roughly 900 participants during August-September 2025. We will stratify the sample in two groups, treatment and control. We will send invitations for the online course on GenAI to the treatment group. The control group will not receive invitations and will not have access to the course through the platform either. We will collect midline and endline surveys roughly 2 and 6 months after the start of the intervention. We will incentivize survey participation with prizes.

Design details

Randomization method
Stratified randomization done in Stata.
Randomization unit
Individual.
Clusters
n/a
Observations
The main sample consists of the roughly 900 individuals for which we have baseline data. Because treatment take up may be low, we will also randomize treatment among the registered users in the platform that did not answer the baseline survey.
Arms
Half of the sample will be assigned to treatment and half will be assigned to control.
Primary outcomes
We will obtain measures on GenAI usage (intensive and extensive margins), prompting sophistication, business practices (marketing, inventory management, record-keeping, and financial planning) and firm performance (revenue and profits).

Cited by

No citation data available for this registration yet.

Keywords