Financial Literacy Coach Pilot Study: User Experience and Behavioral Change Assessment

This study evaluates the impact of MESH Money Coach, an AI-powered financial literacy and credit readiness tool embedded within the MESH platform in Kenya. MESH is a digital platform supporting young entrepreneurs, many of whom cannot access formal credit because they lack basic financial records, documentation, and credit history, despite strong demand. The Coach was built to address this readiness gap at scale, providing personalised financial literacy content, guided habit formation, and credit readiness support across five structured levels. The pilot ran for five weeks with approximately 110 participants recruited from the MESH user base. Participants span a range of ages (18 to 45+), genders, education levels, business stages, and baseline financial readiness levels (Levels 1 to 4 of the Financial Readiness Framework). The intervention was delivered via a conversational AI coach accessible through the MESH platform, with automated nudges, personalised learning content, and open-ended support built in. A baseline survey was administered at Week 0 and an endline survey at Week 5, with 99 baseline respondents and 76 endline respondents, yielding 71 matched participants. These surveys captured financial behaviour across five focal areas: income and expense tracking, savings habits, separation of business and personal finances, financial goal-setting, and credit readiness. Attitudinal measures of financial stress and felt control were collected alongside behavioural indicators. Both the baseline and endline survey data have been collected but not yet fully analysed. Preliminary descriptive comparisons have been produced, but no hypothesis-driven analysis has been conducted. This registration precedes that analysis. Following the quantitative data collection, the study will conduct in-depth individual interviews (IDIs) with approximately 20 participants, selected purposively to represent distinct engagement and progression profiles: early dropouts (Levels 1 and 2), participants who did not change levels, participants who regressed, participants who skipped levels (e.g., Level 2 to Level 4 directly), and those who progressed incrementally. Interviews will be 30 to 45 minutes, conducted remotely or in person, using a semi-structured guide. The interview phase aims to answer what the quantitative data cannot: why certain participants improved sharply, what the lived experience of engagement with the Coach was, what blocked progress for others, and whether observed behavioural shifts reflect genuine habit formation or short-term reporting change. Informed consent will be obtained from all interview participants prior to data collection, and ethical approval and participant consent documentation are being registered alongside this preregistration. The study expects to find meaningful pre-post improvements in financial behaviour scores, particularly in recording frequency and savings consistency, which were the most active targets of the Coach's nudge architecture. The qualitative strand is expected to surface a more complex picture: non-linear progression, context-specific barriers at lower readiness levels, and variation in how different participant groups experienced the Coach's value. The study will also assess whether the Coach generates structured behavioural signals useful for credit eligibility assessment and segmentation, which is a key secondary question for the programme's scale-up phase. No control group was included in this pilot. All findings are directionally attributable to the programme context but cannot rule out selection effects or external influences. This limitation is acknowledged as a structural constraint of the pilot design and will be addressed in any future phase.