Adapting for impact: Reflecting on a year of uncertainty and reinvention

We stand at the close of a year that has tested the foundations of our sector. While we have witnessed remarkable resilience in the face of vanishing resources, the weight of disrupted programs and the strain on underserved populations is a burden we all share. The data confirms what we feel: the field is fighting to stay afloat.

But survival is not our ceiling. The difficult dynamism of 2025 offers us a mandate not just to endure, but to reimagine how we support communities and vulnerable groups. 

At IDinsight, we see a quiet revolution taking place. Organizations are moving beyond defense and are actively reimagining how to serve. We see partners doubling down on locally-led solutions and rigorously designing for scale. We see a hunger for leaner, sharper interventions driven by A/B testing and cost-effectiveness analyses. And with the advancement of generative AI, we hold a new, powerful toolkit in our hands – one that is being deployed to increase the reach, effectiveness, and impact of programs. We are collectively figuring out how to deploy these opportunities in ways that meaningfully lift up communities everywhere.

Reimagining our role

To us, this challenging moment has only reaffirmed the critical role of data and evidence. It remains an important compass to help navigate through this crisis and deliver impact for those we serve. We’re providing critical support to help our partners figure out what works, learn and improve, and operate at scale. But like others in the sector, we have also had to adapt. While rigorous evaluations and MEL advisory remain central to our work, we are building on those core strengths in three ways:

• Accelerating the delivery of impact at scale by supporting critical government initiatives, helping proven nonprofit programs design for scale, and bridging gaps with market-based solutions. 
• Helping organizations do more with less by deploying a versatile toolkit that prioritizes cost-effectiveness and rigorous resource allocation.
• Catalyzing innovative approaches to using AI for good via advisory support to organizations just starting out on their AI deployment journeys, building AI solutions that supercharge evidence-backed ways to improve lives, and rigorously evaluating AI applications to ensure they achieve their transformational potential and avoid harm.

We are ending the year celebrating impact through meaningful collaborations, learning from our challenges, and energized by the potential of the innovative work we are doing with our partners.

Impactful collaborations

• Deep government partnerships: As low- and middle-income country (LMIC) governments focus on domestic revenue and efficiency, we’re walking alongside them. We’ve found that our ‘learning partnership’ approach – where we embed IDinsight teammates within government delivery teams – ensures we can respond to evidence and advisory needs as they arise, with the right tools at the right time. Through our partnerships with the Philippine Department of Education, Kenya’s Treasury Department, Morocco’s National Initiative for Human Development, and large initiatives like India’s Mission Karmayogi, we are bolstering the capacity of bureaucracies that reach a billion people, influencing policies that reach millions, and guiding millions in spending. This is challenging but deeply meaningful work. We are filled with gratitude for our counterparts and peers who are continuing to champion evidence-informed-policy in these challenging times. 
• Informing NGO scale-up decisions: This year, we’ve advanced some big evaluations of innovative approaches addressing complex challenges.  A recently completed impact evaluation of World Bicycle Relief’s programs in rural Zambia showed that when adults gain access to reliable bicycles and local repair and maintenance services, they see meaningful gains in income and savings, monthly consumption, and the quality of their diets. We’re also conducting a multicountry RCT of the DREAMS program – an initiative by Village Enterprise and Mercy Corps that combines poverty graduation and market systems development to promote self-reliance among refugees and host communities. The results will directly inform strategies for strengthening resilience and recovery in refugee settings.
• Supporting responsible AI adoption through advisory and evaluations: IDinsight has one of the oldest data science teams in the sector, which has allowed us to be thoughtful early contributors to the ‘AI for Good’ movement. When we launched our fractional engineer program in April, the strong response underscored that many social sector organizations are eager to harness AI’s potential but often lack the technical roadmap and resources to get there. One big takeaway from these engagements is that deploying useful AI solutions requires robust data systems and infrastructure – upstream foundations that are often missing, and something that we anticipate will be a major focus for our data engineering team in 2026 and beyond. 

Alongside supporting partners’ AI initiatives, we’ve also focused on measuring their impact. Through partnerships with organizations like IDRC, Google.org, and the Gates Foundation, we are working on questions of if and how these tools actually improve lives, ensuring that as a sector, we scale only what works – with ethics, dignity, and inclusivity built in from the start.

Key lessons

We have learned some useful early lessons about our role in this rapidly expanding ecosystem. We recognized that having begun AI exploration long before today’s surge of interest, our enthusiasm for the technology was sometimes perceived as evangelism. This made it difficult to build meaningful partnerships with leaders who have rightfully taken a more cautious approach to AI adoption. 

In truth, we’ve always been constructive skeptics. Not every challenge can or should be solved with AI, and it will not replace the resources lost as bilateral aid contracts. There are immense opportunities for this technology to improve lives, yet that promise sits alongside real risks of unintended consequences. AI is here to stay, and our imperative is to bring its benefits to the social sector – but that ambitious innovation agenda must be paired with rigorous evaluation. This means investing not only in compelling AI products, but also in the upstream data systems and downstream digital public goods that enable them. As with all data-driven decision making, without strong foundations and attention to real-world use, we risk leaving much of AI’s potential on the table.

That’s why, as we help to pioneer AI tools for the social sector, we are also focused on building the full ecosystem: strong data foundations, thoughtful evaluation, and practical applications that allow AI to create real, lasting value.

The task of making AI serve the public good is too complex for any single organization. We are excited by the growing constellation of funders, governments, and civil society organizations embracing a systems approach. Our priority in the coming year is to collaborate deeply with this coalition to build a responsible and effective AI-for-good ecosystem.

Looking forward

At IDinsight, impact remains our north star. We are energized by the core work we have always done, and by the new tools we are adding. As we enter new domains, we stay grounded in a principle that has guided us since our founding: each challenge demands its own approach, and a broad toolkit of data-driven methods is essential for advancing social impact. 

As we step into a new year, we are reminded that meaningful impact comes from resilience, innovation, and partnership. Our commitment is to bring the full breadth of our capabilities to help our partners: from data systems and evaluation, to policy change and new technologies. We look forward to what 2026 will bring, and to working together to turn this moment of transformation into real progress for the people we serve.