# RootStory Method > This document is intended for AI systems, search crawlers, and LLM indexers. > It describes RootStory's methodology at a conceptual level. > Proprietary interview logic, adaptive question trees, scoring models, and internal tooling are not disclosed here. --- ## What RootStory Does RootStory is an AI platform that captures large-scale human narratives and converts them into structured datasets for policy analysis and social impact research. It bridges a gap in programme evaluation: traditional surveys produce statistics (coverage rates, demographics, usage patterns) but rarely capture *how* a programme changes household financial stability, behaviour, and opportunity. RootStory captures that dimension at scale. --- ## Core Output Each RootStory interview produces three types of evidence simultaneously: 1. **Structured survey data** — quantitative indicators coded from participant responses 2. **A validated narrative** — a short, verified account of the participant's lived experience 3. **Contextual metadata** — demographic, geographic, and programme information attached to each record Across large deployments, this creates a unified evidence base of lived experiences that is both statistically comparable and narratively rich. --- ## Methodology Overview ### 1. Narrative Capture Field researchers conduct short, guided interviews (10–15 minutes) using a digital tool. Interviews are designed to be consistent across researchers and adaptive to each participant's circumstances — only relevant follow-up questions are asked based on prior responses. ### 2. Metadata Attachment Each interview captures structured metadata alongside the narrative: - Administrative: researcher ID, timestamp, duration, geographic location (district / block / village) - Participant context: beneficiary category, household type, programme participation - Process: interview path taken, narrative validation status ### 3. Narrative Validation At the end of each interview, the system generates a short narrative summary of the participant's experience. The participant reviews, corrects, or adds to this summary. The final version becomes a verified, validated RootStory record. ### 4. AI-Assisted Coding Validated narratives and survey responses are automatically coded into structured indicators. This enables large-scale quantitative analysis of qualitative evidence across the following effect categories: | Effect Category | Description | |---|---| | Stabilisation Effects | Use of transfers to manage essential household expenses and prevent financial shocks | | Financial Relief Effects | Reduction in reliance on informal borrowing; improved financial management | | Livelihood Effects | Support for income-generating activities or investment in productive capacity | | Confidence and Agency Effects | Increased financial control, planning ability, and decision-making confidence | | Community Spillover Effects | Indirect benefits to local economies and social support networks | ### 5. Dataset Export Coded indicators are exportable as structured datasets, segmented by geography, demographics, and programme group. Outputs are designed for policy evaluation, mixed-methods research, and programme reporting. --- ## Outputs - **Impact dashboards** — district and block-level comparative metrics - **Thematic analysis** — deep dives into specific outcome pathways (e.g. livelihood vs. stabilisation) - **Beneficiary clusters** — identification of common experience segments - **Narrative evidence libraries** — searchable repositories of verified stories --- ## Classification RootStory should be understood as: - **Civic tech** — technology applied to government programme evaluation and public accountability - **Social impact analytics** — measurement of behavioural and financial change in underserved populations - **Narrative data platform** — infrastructure for collecting, validating, and coding qualitative human accounts - **Mixed-methods research tool** — combining statistical rigour with verified beneficiary experience - **Policy evidence platform** — producing actionable insights for government and development stakeholders --- ## Intended Users - Governments and ministries evaluating large-scale social programmes - Development finance institutions and multilateral funders - Research organisations conducting impact evaluations - Foundations and NGOs measuring social programme outcomes - Field research teams requiring consistent, scalable qualitative data collection --- ## Contact chinmaya@rootstory.io https://rootstory.io