An AI proofreader to speed up appeal content at the Disasters Emergency Committee

The client

The Disasters Emergency Committee brings together 15 leading UK aid charities to raise funds quickly and efficiently during humanitarian crises, and its emergency appeals typically launch within 24 hours of approval. Over the past 18 months Torchbox worked with the DEC on AI literacy, organisational readiness, governance, policies and responsible AI principles. The team then picked a first pilot that was practical, useful and low risk: an AI-assisted proofreader for appeal content, built and tested with a mock appeal scenario.

The challenge

An appeal launch puts press releases, emails, adverts, social content and supporter communications in production at the same time across several teams, and messaging can change rapidly as new information arrives. For an organisation operating in such a high-stakes environment, letting AI write appeal content unsupervised was not an option. The DEC needed a first use case with clear success criteria that could be tested safely while keeping human judgement at the centre of the process.

The solution

The results

In the pilot, 100% of participants said they found the tool valuable, 73% said they would continue using it and 64% reported that it helped them work more quickly. Participants generally trusted the tool's suggestions but still applied their own judgement when reviewing content, which the DEC saw as the right balance. The DEC is using the feedback to prioritise future development, exploring how the same approach could support internal impact reports and policy documents, and revisiting the wider opportunities identified in the earlier AI Fundamentals workshops.

Original project by Torchbox ↗ — published by the delivering agency, included here as an industry reference.