By task
Pick the job you are actually trying to do. Every task below is written in the words the organisations we surveyed used themselves.
Translation into and out of local languages
Moving documents between Arabic, Kurdish and English: where machine translation is good enough, and where it is not.
Tools: 6Recipes: 2
Start hereTranscribing testimonies and interviews
Turning recorded interviews and testimonies into text, without the recording ever leaving your organisation.
Tools: 1Recipes: 3
Start hereMaking sense of large document sets
Getting through budgets, court files and hundreds of pages of PDF when nobody has a week to read them.
Tools: 8Recipes: 2
Start hereFollowing legislation, courts and gazettes
Keeping up with the official gazette, parliament and the courts when nothing arrives as an alert.
Tools: 2Recipes: 1
Start hereMonitoring news and social media at scale
Following news and social media at a scale no one person can read, without building a list that puts people at risk.
Tools: 2Recipes: 1
Start hereResearch
Using AI to find leads and starting points, never as a source you can cite.
Tools: 4Recipes: 1
Start hereDrafting reports and advocacy material
Getting a first draft out of a blank page without pasting in what you have not published yet.
Tools: 4Recipes: 2
Start hereCleaning and analyzing data
Making sense of spreadsheets and incident logs, after the identifying columns come out.
Tools: 3Recipes: 1
Start hereVerifying images, video and claims
Checking whether something is real: where AI gives you leads, and where it will confidently mislead you.
Tools: 1Recipes: 3
Start hereGrants and admin
Proposals, budget narratives and donor reports: the paperwork that quietly eats the week.
Tools: 1Recipes: 3
Start hereProtecting sensitive information
Deciding in writing what may go into an AI tool and what never may, before someone has to decide it alone at midnight.
Tools: 2Recipes: 6
Start here