AI is now widely available and seen as innovative across many fields, but not all uses of AI are actually useful for climate resilience or adaptation. What really matters is whether AI adds value to the work organisations are doing on the ground . While local context—such as climate zone or social and economic conditions—shapes the problems being addressed, many AI approaches for analysing environmental and social data are similar across regions. This makes it possible to reuse and scale successful solutions. However, the real impact of AI depends less on the tools themselves and more on whether people, communities, and institutions are willing and able to adopt and use them effectively.

Typical AI use cases within climate resilience

Across all applications, three common AI use cases appear, but most projects combine them rather than using just one. Many follow a simple flow: collecting and monitoring data, forecasting outcomes, and then supporting decisions. These end-to-end AI systems help small organisations gather, analyse, and share information more widely, allowing them to reach more people and better inform communities about issues that affect their everyday lives.

1. Decision Support and Advisory Services for Local Planning: In most applications, the main goal of AI is to turn complex environmental and social data into clear, practical advice for communities and local governments. This information is usually shared through simple tools like dashboards, WhatsApp, SMS, or mobile apps, often in the form of a chatbot. The AI is trained using local data and feedback from communities so it can adapt to local needs and conditions. To ensure accuracy and reduce errors, these systems rely on verified data sources and include human oversight, with most projects building in regular review and feedback. 

Because many organisations work in very remote areas, AI systems need to be adapted to local conditions. This is mainly achieved through:

• ensuring advisory AI tools work offline - meaning the model is already installed on the device (phone, tablet, local computer). Inference (the model produces an output from an input) happens locally.

• deploying edge devices to process data locally near its source to enable faster decisions for IoT and AI (i.e. information is not being sent to costly centralized cloud computing servers); 

• Using Natural Language Processing (NLP) to improve access to information and advice through local languages, specifically in rural communities

2. Monitoring, Data Collection and Detection

Most grassroots organisations start using AI by collecting and combining local and open data, including environmental and social information. They often use on-the-ground sensors to gather accurate local data, while satellite or drone imagery helps fill in gaps and show the bigger picture. AI is then used to monitor this data over time and spot patterns or problems, such as changes in crop health, land use, or infrastructure. Access to European and US satellite data has made this much easier, and many projects also commit to sharing the data they collect, reflecting a strong emphasis on open data.

3. Forecasting & Early Warning Systems

Many applications use AI to help people and organisations understand climate risks and prepare for extreme weather events. AI helps deal with uncertainty by combining past climate data with real-time information from sensors and satellites to predict possible future conditions. By analysing how weather patterns change over time, these models can forecast risks and consider who or what might be affected, taking into account local exposure and vulnerability. To support early warnings, AI presents predictions as probabilities rather than certainties, which is especially useful when data is incomplete and helps provide more reliable, local risk alerts.

Conclusions: the difference is people

Across a thousand applications, clear AI patterns have emerged that can help communities better prepare for climate risks, scale proven solutions, and improve access to information. While these tools strengthen early warnings and decision-making, real innovation is not just technical—it lies in how AI is integrated into human systems. The most impactful projects treat AI as a support for human judgement, local knowledge, and trust, not a replacement for them. AI adds value only when it empowers communities, strengthens local agency, and delivers real-world outcomes. In the end, the success of AI for climate resilience depends less on the technology itself and more on the people that use it.