AI belongs in the workflow. Judgement belongs with you.

The useful question is not simply “What can AI do?” It is “What can it help us do well, with evidence we can trust?” Start there—and build a more responsible way of working.

Colleagues review AI-assisted work against its source evidence

Make responsible use a practical habit.

Define the task

Be clear about the question, the sources available and the decisions the output might inform.

Check the work

Verify claims and calculations. Look for missing context, unsupported certainty and gaps in the evidence.

Own the decision

Protect sensitive information, document limitations and know when a person needs to review the work.

Practise responsible AI within Integrated MEL.

The AI foundations course helps you understand the tools and their limits. The wider programme connects responsible AI with monitoring, evaluation, accountability, learning and budgeting.

Explore Integrated MEL

Explore our AI and data learning topics

These pages outline additional learning topics. They are previews, not separately available courses or enrolment offers. For a programme you can join now, explore Integrated MEL.

AI for NGOs: Use It Safely, Use It Well
Topic preview

AI for NGOs: Use It Safely, Use It Well

Explore how AI can support programme work, where it falls short and what responsible use asks of your team.

Explore this topic
AI Foundations: Verification is Mandatory
Topic preview

AI Foundations: Verification is Mandatory

Build the habit of checking sources, numbers and claims before an AI-assisted draft becomes programme evidence.

Explore this topic
Prompting and Workflow Skills for MEL
Topic preview

Prompting and Workflow Skills for MEL

Give AI a clear task, relevant context and a useful output format, then build review into the workflow.

Explore this topic
Anatomy of a Working Prompt
Topic preview

Anatomy of a Working Prompt

Turn a vague request into a structured prompt with a clear purpose, source material, boundaries and review criteria.

Explore this topic
The Data Sharing Ladder
Topic preview

The Data Sharing Ladder

Pause before sharing programme data. Consider its sensitivity, the permissions you have and the environment you are using.

Explore this topic
The Data Check Habit and the First Hour
Topic preview

The Data Check Habit and the First Hour

Begin with data checks that matter: definitions, completeness, duplicates and the context behind an unexpected result.

Explore this topic
Subscription vs Per-Token: AI Costs Explained
Topic preview

Subscription vs Per-Token: AI Costs Explained

Compare the costs behind AI access, including usage, staff time, verification and support.

Explore this topic
API Basics for Non-Developers
Topic preview

API Basics for Non-Developers

Understand how tools exchange information and the questions to ask before connecting an AI service to a MEL workflow.

Explore this topic
Running Models Locally
Topic preview

Running Models Locally

Explore what local AI could mean for your team, including hardware, maintenance and responsibility for data.

Explore this topic
Offline-First Field Setups
Topic preview

Offline-First Field Setups

Think through fieldwork when connectivity is unreliable: collection, local storage, safe transfer and synchronisation.

Explore this topic
The Open Source Stack for MEL Data
Topic preview

The Open Source Stack for MEL Data

Consider how open-source tools fit together and what your team needs to maintain a reliable evidence workflow.

Explore this topic
Hugging Face: The Model Library
Topic preview

Hugging Face: The Model Library

Learn what to look for in a model library, including documentation, intended use, limitations and licence conditions.

Explore this topic
The Do More With Less Trap
Topic preview

The Do More With Less Trap

Challenge the idea that faster output automatically means better work. Make quality, care and realistic workloads part of the discussion.

Explore this topic