Research, narrative intelligence, and the tools between them.

I'm Senior Officer for Communications Monitoring and Analysis at the International Federation of Red Cross and Red Crescent Societies (IFRC). My practice covers monitoring and analysis across the world's largest humanitarian network, social listening in 30+ languages, media monitoring, and executive-level analytics, with a specialisation that has become the centre of my work: harmful information. Impersonation and donation-scam campaigns during disasters. Health disinformation during outbreaks. Narratives that put field staff at risk.

The monitoring tools this work needs mostly don't exist off the shelf, so I taught myself to build them. I build and run three automated detection-and-triage pipelines, Python, the Claude API, GitHub Actions, that score multi-source signal against a harm taxonomy, tier sources by risk, and ship daily assessments to an opt-in expert audience, currently through a live public-health emergency. I built and operate them independently, and most of what I know about keeping an LLM system honest I learned from their failure modes. The case study tells that story properly.

I also work the institutional side of the same problem: I am part of the team drafting the Red Cross Red Crescent Movement's 2026 Council of Delegates resolution on harmful information, developed through consultation with 191 National Societies, and I supported the July 2026 launch of the Movement Hub on Harmful Information, the Movement's joint body on the issue. I contributed to the harmful-information workstream around the World Disasters Report 2026.

Before that, I spent four years at UNICEF headquarters on digital fundraising integration and analytics, dashboards and automated reporting across HQ and 33 National Committees, cross-domain tracking architecture, and rapid-response monitoring during emergency campaigns. Before that: startup years, including marketing communications at a fintech and co-founding a small digital consultancy. I'm French and American, and work in French and English.

I am usually most useful when the problem is messy: too much information, no shared definition of what matters, and a decision that still needs to be made. I like building the bridge from signal to judgment, whether that takes the form of a research brief, a measurement framework, a policy recommendation or a Python workflow.

ContactThe fastest way to reach me is the contact form, lucile.tranchant@gmail.com, or LinkedIn.