Opinion: Development organizations need an action plan for AI adoption

Development organizations face unprecedented challenges. The uncertain environment has broad ramifications for development funding and programming, with the recent USAID cuts threatening to strip 23 million children of educational access and 95 million people of basic health care. In order to navigate this turbulence and address the gap, organizations need to deliver more impact with drastically reduced resources. There’s a powerful solution hiding in plain sight: Artificial intelligence tools that can dramatically amplify organizational capacity. A Harvard Business School study found that consultants using AI had significant increases in productivity, completing 12% more tasks, 25% faster, and with a 40% increase in quality. Still, 92% of nonprofits feel unprepared for AI implementation. This skills gap represents both the sector’s greatest vulnerability and its most immediate opportunity. The productivity crisis demands bold action The math is unforgiving, yet the current reality holds both challenge and opportunity. Development organizations must deliver more impact with shrinking resources, yet many remain trapped in labor-intensive processes that consume precious time and talent. Grant writing, donor reporting, program monitoring, and communications — core functions that drive organizational survival — still rely heavily on manual, repetitive tasks that AI can streamline dramatically. Consider the current paradox: Over half of nonprofit staff surveyed across New York State use AI at work, despite only 11% receiving guidelines on AI use. This shadow AI usage — i.e., the unauthorized or unmanaged use of AI tools within an organization — is attracting the notice of auditors due to data privacy concerns and governance implications. Yet even without guidance or training, results are emerging, with 30% of nonprofits using AI achieving boosted fundraising revenue. Meanwhile, the for-profit sector is racing ahead — two-thirds of companies are actively implementing at least 10 AI pilots. Organizations dedicated to solving the world’s most pressing challenges are themselves being left behind by the technological advances that could amplify their impact. Here four areas for development organizations to consider for successful AI adoption: 1. A human-centered approach to AI adoption The path forward requires addressing AI adoption fundamentally as a people-and-process challenge, not a technology one. Research across industries shows that 70% of AI implementation obstacles stem from people and process issues, with only 10% related to the underlying algorithms. At ImpactAgent, the organization I founded to help mission-driven organizations improve their organizational effectiveness with AI, we focus on the human elements that determine success. 2. AI adoption must come from both leadership and the workforce Organizations risk being stuck in “pilotitis” if they don’t bring their entire workforce along and tailor their AI adoption to the capabilities of their workforce. Effective AI adoption can’t be either a purely top-down mandate nor bottoms-up experimentation, but rather a synchronized approach. Research shows that managerial support is the strongest predictor of AI usage, while network effects are also crucial — workers are three times more likely to use AI if they know others using AI. 3. Implement practical governance frameworks Effective AI adoption requires clear but practical governance, especially in the light of current shadow AI usage. ImpactAgent helps organizations establish guidelines around data privacy, accuracy standards, and appropriate use cases. If an organization doesn’t yet have an AI policy but has high shadow AI usage, we suggest it implement a simple AI policy to start and revisit over time as AI becomes more embedded in its workforce and workflows. 4. Start with targeted use cases Successful AI adoption begins by focusing on specific, well-defined use cases. AI excels at tasks involving repetitive processes, data analysis, and content generation — areas where development organizations face efficiency challenges. Identifying key pain points in an organization’s existing workflows and targeting AI solutions to help address them will help. Good starting points to explore include streamlining grant writing and reporting, enhancing donor engagement, and automating monitoring and evaluation. The goal is not to replace human judgment, but rather to amplify human capacity for the work that matters most: building relationships, solving complex problems, and creating lasting change in communities worldwide. Time for action The development sector stands at a crossroads. Organizations that embrace AI strategically and responsibly will unlock unprecedented capacity to advance their missions. Those that hesitate risk being left behind as funding pressures intensify and donor expectations evolve. The path forward requires immediate action: upskill workforce through practical training, establish governance frameworks, and launch targeted pilots. The technology exists. The need is urgent. The communities we serve — and the challenges they face — cannot wait for perfect conditions. They need development organizations operating at peak efficiency, using every available tool to maximize impact. AI is a critical tool development organizations should utilize in a world where resources are scarce but human need is vast. Author’s note: For full disclosure, this opinion piece was written with the assistance of Claude, to which I fed my detailed prompt, guidance on tone of voice and background documents. I then manually edited the draft, cross-checked all the references, and also checked the output using ChatGPT. I estimate usage of AI saved me approximately five hours in the creation of this piece.