Relief Applications

Glossary of data, ERP, MEAL and AI terms for NGOs

The words used around data protection, management systems, monitoring and AI are often borrowed from law, software and statistics, and they are not always explained. This glossary gives each one in a sentence or two, in the way NGO teams use them, and links to the guide that goes deeper.

The short answer

This glossary defines the terms NGO staff meet when they work on data protection, ERP and operations systems, monitoring, evaluation, accountability and learning (MEAL), and AI. Each term is explained in plain words, and the longer guides on this site are linked where they help. The meanings are general. Where a law or a donor sets its own definition, that definition is the one that counts.

Data protection terms

These terms come mostly from the GDPR, the European data protection law, which many NGOs follow even outside Europe because a donor or a partner asks them to. Our guides to whether an NGO needs a DPO and to DPIAs for NGOs use them in detail.

Anonymisation
Changing data so that nobody can identify a person from it, by any reasonable means. Data that is truly anonymous falls outside most data protection law. Removing names alone is usually not enough
Consent
A person’s free, specific, informed and clear agreement to the use of their data. It is only one of several lawful bases, and it is hard to rely on where people depend on you for help
Controller
The organisation that decides why and how personal data is used. An NGO is the controller for the data it collects about the people it serves and about its staff
Data breach
A security incident that leads to the loss, destruction, alteration or unauthorised disclosure of, or access to, personal data
Data minimisation
Collecting and keeping only the personal data you need for a stated purpose, and no more
Data protection impact assessment (DPIA)
A structured review of the risks to people from a use of their data, done before the work starts, and the steps to reduce them See DPIAs for NGOs.
Data protection officer (DPO)
A person who advises an organisation on data protection and monitors its compliance. Some organisations must appoint one, and others do so by choice See does an NGO need a DPO.
Data subject
The person whom personal data is about, such as a beneficiary, a member of staff or a donor
GDPR
The General Data Protection Regulation, the European Union’s data protection law, which applies to organisations in the EU and, in some cases, to those outside it that handle the data of people in the EU
Lawful basis
The legal reason that allows you to use personal data, such as consent, a contract, a legal duty, protecting someone’s vital interests, a task in the public interest or a legitimate interest
Personal data
Any information about an identified or identifiable person, such as a name, a phone number, a photo, a location, a case number or an ID card number
Processor
An organisation that uses personal data on behalf of a controller and on its instructions, such as a software provider that hosts your data
Pseudonymisation
Replacing identifying details with a code, with the key kept apart. The data is still personal data, because the key can bring identification back
Sensitive data
Data that carries a higher risk to people if it is misused, such as health, ethnic origin, religion, political opinion or sexual orientation. The GDPR calls it special category data and gives it stronger protection

ERP and operations terms

An ERP brings the work of an organisation together on one base. Our ERP buyer’s guide for NGOs explains what to look for, and the guides on case management, cash and voucher assistance and a complaints and feedback mechanism cover three common needs.

API
Application programming interface: a defined way for one piece of software to ask another for data or send it data. It is how systems are connected without retyping
Audit trail
A record of who did what and when in a system, so a change or an approval can be traced
Case management
Following an individual or a household through a service, from the first contact through assessment, action and closure, with a record of each step See humanitarian case management systems.
Cash and voucher assistance (CVA)
Help given as money or vouchers instead of goods, so that people choose what they need. A CVA system registers people, checks eligibility, makes the payments and reconciles them See cash and voucher assistance software.
Complaints and feedback mechanism (CFM)
The channels and the process through which the people an organisation serves can raise a concern or give a comment, and get a response See how to set one up.
ERP
Enterprise resource planning: software that runs the main functions of an organisation, such as finance, grants, human resources, procurement and projects, on one shared base See ERP for NGOs and Relief One.
Grant management
Tracking funding from a donor from the proposal through budgets, spending, reports and closure
Integration
A connection between two systems so that data moves between them without being typed again
Module
One functional part of an ERP, such as finance or fleet. You pick the modules you need and add others later
Offline-first
Software that works without a connection and sends its data when the connection returns, which matters in the field
Role-based access
Giving each person the access their role needs, and no more, so that a user sees only the records they should

MEAL and information management terms

MEAL stands for monitoring, evaluation, accountability and learning. Our guides on monitoring and evaluation tools and on building a Power BI dashboard from KoboToolbox data put these terms to work.

Baseline
The value of an indicator before a programme starts, which later values are compared with
Dashboard
A page of charts, maps and figures that shows the state of a programme at a glance, usually fed from live data
Enumerator
A person who collects survey data by asking questions and recording the answers
Evaluation
A structured judgement of whether a programme worked, why, and what to learn from it, usually at the middle or the end
Indicator
A measure used to track progress, such as the number of people reached or the share of households with safe water
Information management
Collecting, storing, checking, analysing and sharing data so that people can make decisions with it See our information management service.
KoboToolbox
A free, open-source tool for collecting survey data on phones and in the browser, widely used in humanitarian and development work See KoboToolbox to Power BI.
Logframe
Logical framework: a table that sets out a project’s goal, outcomes, outputs and activities, with the indicators, the means of verification and the assumptions for each
MEAL
Monitoring, evaluation, accountability and learning: the cycle of tracking a programme, judging it, answering to the people it serves and using what is learned See monitoring and evaluation tools.
Monitoring
Regularly collecting and checking data while a programme runs, to see whether it is on track
Power BI
Microsoft’s business intelligence tool, which connects to many data sources and builds dashboards See KoboToolbox to Power BI.
Target
The value an indicator is meant to reach by a set date
XLSForm
A standard for writing forms in a spreadsheet, used by KoboToolbox and other data collection tools

AI terms

These terms matter when you write rules for AI or build a chatbot. See our guides to an AI policy for NGOs and to humanitarian chatbots.

AI governance
The rules, roles and checks that decide how an organisation uses AI and who answers for it See AI policy for NGOs.
Chatbot
A program that talks with people in a chat window or a messaging app. It may follow fixed rules, use a language model, or both See humanitarian chatbots.
Guardrails
The limits placed on an AI tool to keep its output within what is safe and intended, such as prompt rules, filters and human review See humanitarian chatbots.
Hallucination
An answer from a language model that sounds confident and is wrong or made up
Human in the loop
A design where a person reviews or approves what an AI produces before it has an effect on anyone
Language model
An AI system trained on large amounts of text that writes replies to questions. A large language model, or LLM, is one at a big scale
Prompt
The instruction or question given to a language model. A system prompt, written by whoever sets up the tool, sets the limits on what it may answer. Writing prompts well is called prompt engineering
Rule-based chatbot
A chatbot whose answers are all written in advance and chosen by menus, buttons or keywords
Training data
The material an AI system learns from. Ask a vendor whether your own data is used to train its models See AI policy for NGOs.
Transparency
Being open about where and how AI is used, including telling people when they are talking to an AI system

How do you use this glossary?

Use it to agree on words before a project starts. Many misunderstandings between programme, IT and legal teams come from a term that means something different to each of them. “Anonymous” is a good example: to a survey team it means no name on the form, and to a lawyer it means that nobody can be identified by any reasonable means.

If a donor or a law defines a term differently, use that definition in that setting. If you cannot find a term here, ask us, and we will add it.

Questions and answers

What is the difference between a controller and a processor?

A controller decides why and how personal data is used. A processor uses it on the controller’s behalf and on its instructions. An NGO is usually the controller, and a software host is usually a processor.

What is the difference between anonymised and pseudonymised data?

Anonymised data cannot be tied to a person by any reasonable means. Pseudonymised data replaces identifying details with a code, and the key kept apart can bring identification back, so it is still personal data.

What does MEAL stand for?

Monitoring, evaluation, accountability and learning. It is the cycle of tracking a programme, judging it, answering to the people it serves and using what is learned.

What is an ERP for an NGO?

Software that runs the main functions of the organisation, such as finance, grants, human resources, procurement and projects, on one shared base.

What is a hallucination in AI?

An answer from a language model that sounds confident and is wrong or made up. Limits on what the model may say, approved content and human review reduce the risk.

A term missing, or a question about one?

Tell us which word you would like explained, or which project you are trying to describe, and we will help.

Keep reading

This guide is general information, not legal or technical advice. Reviewed on 11 November 2026 by the team at Relief Applications. Definitions are general. A law or a donor may define a term differently.