Three series covering the work of funding and building AI in mission-driven, low-resource contexts: from funder due diligence to data architecture to portfolio support.
A funder's toolkit for backing AI in mission-driven, low-resource contexts. Frameworks for readiness assessment, evaluation criteria, scale strategy, and documented case studies of what works and what causes harm.
A four-dimension, four-stage maturity model for portfolio-level investment decisions — with stage-specific playbooks for what to fund and what not to fund, and self-assessments for both your portfolio organisations and your own internal team.
Read the framework → Field GuideMost foundations are funding AI enthusiasm. The ones doing it well are funding AI evidence. Five questions that tell you which is which — covering workflow clarity, data governance, evaluation design, government adoption, and failure modes.
Read the field guide → Field GuideScale is used to mean a dozen different things. This field guide maps three distinct types — Steady Impact, Linear Growth, and Exponential Impact — what each actually requires, and where AI has begun to change the geometry of what's possible.
Read the field guide → Case StudiesThree deployments that worked. Three that caused documented harm. The structural patterns that distinguish them — drawn from twenty government-supported AI deployments in India between 2023 and 2026.
Read the case studies →Who owns the system controls the intervention. A growing series on the data design decisions that determine whether an AI deployment serves communities or extracts from them.
A four-question framework and four-tier ownership model for designing data architecture deliberately — before the first commit, not after the first failure.
Read the field guide → Field Guide · Part 2Most social sector tech is built as a single application that does everything. This works until requirements change, a new language is needed, or the original developer moves on. A field guide on building for adaptability.
Read the field guide → Field Guide · Part 3Packaging content inside the application is fast at first and expensive forever. How treating content as structured data — atomic, language-tagged, configuration-driven — changes what becomes possible.
Read the field guide → Field Guide · Part 4Three localization models — manual, fully automated, and human-in-the-loop — and why only one of them works at the language diversity and quality bar the Global South actually requires.
Read the field guide → Field Guide · Part 5Engagement metrics are easy to collect and easy to manipulate. How to design data architecture that captures learning outcomes, behavioral change, and systemic trust — not just usage.
Read the field guide →How funders, accelerators, and foundations can build support infrastructure that actually reaches portfolio organizations — from what VC firms got right to mentor network design to AI opportunity sequencing.
The VC model is worth studying. What transfers, what breaks, and the insight the VC playbook can't give you about building community in mission-driven portfolios.
Read the field guide → Field Guide · Part 2Fundraising strategy, AI implementation, government partnerships, product scaling, peer community. A diagnostic framework for matching support to the real need, not the stated one.
Read the field guide → Field Guide · Part 3Most mentor networks fail because they were built as directories instead of systems. Four design principles — segmentation, capability-based recruiting, diagnosis-led matching, and structured cadence.
Read the field guide → Field Guide · Part 4The gap between "AI-powered" in a pitch deck and AI working in production is where most funder energy disappears. How to help your portfolio sequence the investment correctly.
Read the field guide →Working with funders and foundations on AI strategy. To discuss how these frameworks apply to your portfolio, write to shruti@tiltedground.com or book a call.
Book a free call →