America Succeeds used Evidano in a four-site study of how schools build durable skills

America Succeeds and researchers from Northeastern University asked how schools can deliberately help students develop durable skills such as communication, critical thinking, collaboration, creativity, leadership, metacognition, and adaptability.

Phase I studied three high schools and one profession-based program: Bostonia Global High School in California, Gibson Ek High School in Washington, Building 21 Philadelphia in Pennsylvania, and GO CAPS Monett in Missouri. The team examined what each program did, how students experienced it, what counted as evidence of skill growth, and which structures made the work possible. The full report is published by America Succeeds.

The project began with two online leadership interviews at each site. Researchers then conducted two-day visits in spring 2024, observing daily practice and holding semi-structured focus groups and interviews with educators, students, staff, and, where relevant, industry partners.

Five alumni, four from GO CAPS Monett and one from Gibson Ek, described how program experiences transferred into post-secondary education and work. Survey pilots at Gibson Ek and GO CAPS Monett added quantitative and open-ended evidence about durable skills, self-reflection, and program practices.

This design let the team compare leadership intent, observed practice, current student experience, alumni reflection, and survey evidence instead of relying on one source alone.

The report states that Evidano, then AILYZE, ChatGPT, and Claude were used to conduct the initial analysis of all interview and focus-group transcripts. The tools helped the team identify recurring themes, core practices, cross-stakeholder patterns, and connections between program practices and reported outcomes.

The AI-assisted first pass was followed by a manual validation process. Researchers checked every AI-generated analysis against original transcripts and video recordings. Draft case studies were then shared with school and program leaders for accuracy review. Finally, a systematic cross-site comparison distinguished common practices from local adaptations.

Across different school types and communities, the report identified five recurring practices: interest-driven and project-based learning; real-world experiences such as internships, mentoring, and client projects; competency-based assessment; digital platforms and portfolios that make growth visible; and intentional advising and coaching.

The report argues that these practices reinforce one another. Projects give students a reason to apply skills; real-world settings make the stakes authentic; competency frameworks clarify expectations; portfolios support reflection; and advisers help students interpret feedback and plan their next steps.

The report also emphasizes adaptation. Bostonia Global, Gibson Ek, Building 21, and GO CAPS Monett used different schedules, technologies, partnerships, and assessment systems. The shared lesson was not to copy one model exactly, but to align curriculum, experience, evidence, and adult support around explicit skills.

This case shows how AI-assisted qualitative analysis can sit inside a multi-site research design without becoming the sole analytic authority. Evidano helped the team process a broad transcript corpus and locate patterns across stakeholder groups. Human researchers then verified the output against source media, and participating institutions checked the resulting narratives.

It is also a useful example for foundations, education nonprofits, and research-practice partnerships that need to turn large volumes of site evidence into individual case studies and a cross-case synthesis on a practical timeline.

The report is an institutional research publication rather than a peer-reviewed journal article. The four sites were purposively selected because they had explicit commitments to durable-skill development, so the findings should not be read as a representative estimate of all U.S. schools.

The analysis used three AI systems. The report discloses Evidano as part of the initial analysis but does not separate the output or performance of each platform.