Software

Also called: program, application.

Detail

In five seconds

What it is

A set of coded instructions that tells a computer how to take input, process it and produce output.

In plain words

Software is the set of instructions that makes a computer or phone do useful tasks, from showing a webpage to checking records.

An everyday example

In a city registry office, a clerk receives forms from many counters. The office manager gives the clerk a written checklist: read the name, compare birth dates, flag possible duplicates and place each form in the correct tray. The same desk and trays stay the same, but the checklist can be updated when the office changes its procedure. At the end, the office produces a clean list and a small stack of records that need human review.

Where the story stops working

A clerk can use judgement, ask questions and handle unusual cases in flexible ways. Ordinary software does only what its coded instructions and connected components allow. Also, real software can run millions of steps per second and may include many modules, which a single office checklist does not show.

Remember thisSoftware is the whole working application, including code, interfaces, configuration and supporting components. Algorithms and AI models may sit inside it, but they are not the whole system.

This term is contested

The core idea of software is not very disputed, but the boundary between ordinary software and AI software is often debated, especially in marketing and policy discussions.

Understand it

Why it matters

Non-specialists hear many digital terms used as if they mean the same thing. Knowing what software is helps you ask better questions: What data goes in? What steps are applied? What output comes out? Is the problem in the hardware, the software, the data or the human process around it? This matters for procurement, risk, accountability and everyday use. A vendor may describe a product as AI when it is mostly ordinary software with fixed rules. A useful discussion depends on separating the overall application from one algorithm or one model inside it. That helps people judge reliability, privacy, cost, maintenance needs and where human oversight is required.

How software turns input into output

After viewing this, a reader should be able to explain that software sits between input and output, applying coded instructions on hardware.

01
Input arrives
Data or user actions enter the system, such as typed names, button clicks or uploaded files.
Draw a box labelled 'Input' with icons for text, numbers and a mouse click. Add example caption: 'registry records'.
02
Software gives instructions
The program contains coded rules telling the computer what steps to perform.
Draw a box labelled 'Software' with stacked lines representing code. Add a small note underneath: 'instructions, not the machine itself'.
03
Processing happens
The software checks, sorts, compares, calculates or transforms the input.
Draw a box labelled 'Processing' with three small tags inside: 'check', 'compare', 'sort'. Show arrows moving through the tags.
04
Output appears
The system returns a result, such as a report, alert, saved record or updated screen.
Draw a box labelled 'Output' with icons for a report, warning sign and saved file. Add a side note: 'may still need human review'.

A left-to-right flow diagram with four boxes connected by arrows. Read from input to output. The second box names software as instructions, and the third shows processing actions such as checking, sorting and calculating. Small labels below show that hardware runs the software but is not the same thing.

Hovering or tapping each box reveals one sentence with an example: input record, coded rule, duplicate check, flagged result.

Describe this without the diagram

The diagram is a four-step flow from left to right. First, input enters the system as text, numbers or clicks. Second, software is shown as coded instructions that tell the computer what to do. Third, processing applies those instructions to check, compare, sort or calculate. Fourth, output appears as a report, alert or saved file. A note explains that hardware runs the software, but hardware and software are different parts of the system.

With reduced motion: A static four-box flow replaces any animation. Arrows show the same sequence: Input, Software, Processing, Output, with the same labels and notes visible at once.

Software is the part of a digital system that tells the hardware what to do. Hardware is the physical machine: the screen, processor, keyboard, storage and network equipment. Software is the coded instructions and rules those parts follow.

Something goes in, such as text, numbers, clicks, images or data from another system. The software applies steps to that input. Those steps might be simple, like sorting names, or more complex, like checking for duplicate records or generating draft text.

Then something comes out. The output might be a message on screen, a saved file, a warning, a chart or an action in another system. If the software has a fault, often called a bug, the output may be wrong even when the computer itself is working normally.

Software is a broad term. An algorithm is a method or procedure; software is the full working implementation of one or more methods in code. A model, in AI, is usually one component inside software, not the whole application. So not all software uses AI, and not all parts of AI systems are models.

How “A registry office checklist and clerk” maps onto the real thing

In the story	In the AI system
the written checklist	the software instructions
forms arriving from counters	input data
sorting into trays and flagging duplicates	processing steps and output
the desk and trays	computer hardware

A worked example

The statistician prepares the programme status table that will support the talking points.

Starting from: A workbook containing reporting-period, location and activity-status fields, with one blank status and one location written in two different ways.

The software receives the workbook through its upload interface.
A configuration file identifies the required fields and accepted status labels.
Validation logic flags the blank status and standardises the two location labels against an approved reference list.
A database query groups the validated rows, and the export component places the summary and exceptions in separate tables.

Result: The software produces a summary table for the talking points and an exceptions table for follow-up. No AI model is required for these operations.

In a United Nations workplace

Ahead of informals, a statistician in a country office is asked by the regional bureau for talking points by close of business. The statistician uses an approved reporting application to import implementing partner returns, check required fields and produce a summary table for clearance.

The application is software: its interface accepts the files, validation logic checks them, configuration rules define accepted fields, database queries assemble the results and an export function produces the table. It may do all of this without using an AI model.

Before you try thisOnly data authorised for that application may be loaded. Personal or protection-sensitive fields that are not required must be removed, and such material must never be transferred to an unapproved tool because its access, retention and reuse controls may be unsuitable.

Where you may hear it

During clearance of terms of reference for a new information system
At a vendor demonstration in a procurement meeting
In an ICT governance board discussion on approved applications
When the front office asks whether a proposed AI feature requires new software

Avoid confusion

Often confused with

Common misconceptions

Common misconception: Software means the same thing as AI.
Correction: AI is only one subset of software. Many programs follow fixed rules and do not learn from data. Calling all software AI can hide what the system actually does and what risks it creates.
Common misconception: If software gives a result, the result must be correct.
Correction: Software can fail because of coding bugs, poor design, wrong settings or bad input data. A computer can follow instructions exactly and still produce an unhelpful or incorrect output if the instructions or data are flawed.
Common misconception: An algorithm and software are the same thing.
Correction: An algorithm is a method for solving a problem, like a procedure on paper. Software is the implemented code, interface, error handling, connections and other parts needed to run that method on a real machine.

Quick check

Which statement best describes software?

Go deeper

The simple version

Software is the set of instructions that makes a computer or phone do useful tasks, from showing a webpage to checking records.

The technical version

Software is executable code and related components that run on hardware to perform specified operations on inputs and produce outputs. In practice, it often includes program logic, user interface elements, configuration files, libraries, database queries, network calls and error handling. One software application may contain many algorithms and, in some systems, one or more AI models. That is why software is broader than an algorithm and broader than a model. The same hardware can run different software, and changing the software can change the system's behaviour without changing the physical machine.

Reviewed external resources

Primary sourceRecommendation on the Ethics of Artificial Intelligence (opens in a new tab)UNESCO
This link gives a public-sector ethics frame for digital and AI systems and suits readers who need governance context rather than a technical definition.
Intermediatelink checked 2026-07-31
EncyclopaediaSoftware (opens in a new tab)Wikipedia
This link gives a broad introductory overview of software and suits readers who want a quick general definition with historical context.
Intermediatelink checked 2026-07-31
ArticleMachine Learning Glossary (opens in a new tab)Google
This link gives short plain-language explanations of AI-related terms and suits readers who want to see how software differs from machine-learning vocabulary.
Intermediatelink checked 2026-07-31
VideoVideo explanations: Software (opens in a new tab)YouTube search
This search results page offers many video explainers on software and suits readers who prefer to compare several teaching styles before choosing one.
search results, not one pagelink checked 2026-07-31
DiscussionCommunity discussion: Software (opens in a new tab)Reddit search
This search results page shows community discussions about software and suits readers who want informal questions, examples and points of confusion people commonly raise.
search results, not one pagelink checked 2026-07-31

Where this came from

Drafted with a language model through an offline pipeline, checked against this site's schema, and every external link above fetched to see whether it answers. Treat it as a starting point and check anything that matters against the primary sources. How this site is built and what that means.

The one thing to remember

Software is the whole working application, including code, interfaces, configuration and supporting components. Algorithms and AI models may sit inside it, but they are not the whole system.