Prosper Global deploys AI agents to accelerate humanitarian crisis research

What's Happening

• Prosper Global partnered with Cloudera to build VERA, a multi-agent research engine that accelerates crisis response analysis.
• The system runs on Amazon Web Services and uses Anthropic models to synthesize data from conflict tracker ACLED and local news.
• In Sudan, the tool reduced agricultural desk research from six days to two, saving about $1,500.

Why It Matters

Agentic AI is entering high-stakes humanitarian environments. By automating the initial data gathering phase, NGOs deploy resources faster while keeping human experts in control of final assessments.

The Context

• Prosper Global, formerly known as Mercy Corps, operates in 30 countries facing conflict, climate change and economic hardship.
• Traditional crisis research cycles take days to weeks of manual data collection before analysts draft reports.
• Prosper Global operates VERA, a system launched in June, across Ukraine, Colombia, and the Democratic Republic of Congo.

What They're Saying

• The system provides a starting point for analysts rather than a finished product, Prosper Global director of data science Aaron Eubank said.
• "We want to keep the agency in the hands of the researchers themselves," Eubank noted.

What to Watch

• Prosper Global plans to add more resources to the agentic system to create a curated research process for future crises.

The Bottom Line

Humanitarian organizations are adopting multi-agent AI to process conflict data faster without removing human oversight from critical decisions.