Learn AI /AI Foundations /Automation
Automation
- Starter
- AI Foundations
- Policy officers
- Managers & front offices
Also called: automated system.
Back to the milestone you came from
In five seconds
What it is
Automation uses a machine or software to carry out set steps with little or no human action after setup.
In plain words
Automation means getting a machine or software to do repeated steps automatically after people set up the process.
An everyday example
A phone can be set to back up new photos whenever it is charging and connected to Wi-Fi. Once those conditions are met, it identifies files not yet copied, uploads them and reports whether the backup succeeded. The owner does not start each transfer manually, but still needs to notice failed uploads and decide which photos should be stored.
This analogy focuses on clear, rule-based checking. Real automation can handle many other tasks, including moving files, sending messages, or controlling equipment. Also, some automated systems use machine learning for messy inputs such as images or free text, while the warehouse example mostly relies on exact matches and clear rules.
Remember thisAutomation carries out a defined workflow with less direct human action. AI may be used in one step, but automation alone does not provide judgement, accuracy or accountability.
This term is contested
The basic idea is widely agreed, but the boundary between simple automation and AI-enabled automation is often described differently by different organisations and vendors.
Understand it
Understand it
Why it matters
Automation affects budgets, speed, staffing, risk and accountability. If you hear that a process is being automated, you need to ask basic questions: what task is being automated, what rules or model does it use, what data goes in, what errors can happen, who checks exceptions, and who is responsible for the outcome. This matters in public services and international organisations because automation can improve consistency and free staff time, but it can also make unfair, unsafe or low-quality results spread faster if the design is poor or oversight is weak.
How automation turns an input into an output
After viewing this diagram, a reader should be able to explain that automation takes an input, applies set steps, and produces an output, sometimes with a human review point.
01
1. Input arrives
The system receives something to work on, such as a form, barcode, email, payment record or sensor reading.
Draw a rounded rectangle labelled 'Input' with small icons for a document, barcode and data table inside or below it.
02
2. Steps are applied
Software follows instructions or rules. For example, compare values, sort items, move a file or trigger an alert.
Draw a process box labelled 'Rules or process' with a simple checklist icon. Connect Input to this box with an arrow.
03
3. Result is produced
The system creates an output such as a report, approved transaction, updated record or warning message.
Draw a box labelled 'Output' with icons for a report and alert. Connect the process box to this box with an arrow.
04
4. Human checks exceptions
If the system finds something unusual or high-risk, a person reviews it before action is final.
Place a smaller box slightly below the Output box labelled 'Human check when needed'. Use a branch arrow from the process or output stage to this review box, then an arrow back to Output.
A left-to-right flow diagram with four boxes connected by arrows: Input, Rules or process, Output, and Human check when needed. Read from left to right to see how a task is carried out automatically and where a person may step in.
Hover or tap each box to reveal one short example: invoice processing, manifest checking, access control, and survey coding.
Describe this without the diagram
The diagram is a left-to-right flow with four stages. First, an input arrives, such as a form, barcode, email or sensor reading. Second, software applies set steps or rules, such as comparing values, sorting items, moving files or creating an alert. Third, the system produces an output, such as a report, updated record, approval or warning. Fourth, if something unusual or high-risk appears, a human checks that case before the action is final. The main idea is that automation does repeated work automatically, but people may still review exceptions.
With reduced motion: A static four-box flow chart shows Input, Rules or process, Output, and Human check when needed, with arrows showing the main path and the exception path.
Automation means setting up a process so a machine or software carries out repeated steps on its own. The input might be data, a document, a sensor reading, or a button press. The process follows instructions. The output might be a sorted file, an alert, a payment, or a message.
Not all automation uses artificial intelligence. A simple automated system may follow fixed rules such as “if the amount is above this limit, send it for review”. That is different from machine learning, where a model finds patterns from examples and then makes predictions or classifications.
What goes in matters. If the input is wrong, incomplete or delayed, the output can also be wrong. This is sometimes called “garbage in, garbage out”. Automation can make work faster and more consistent, but it can also spread mistakes quickly if no one checks the process.
A useful way to think about automation is: people design the steps, the system executes them, and people remain responsible for the result. In some cases a human reviews or approves key points. That is often safer than letting a process run without oversight.
How “A phone backing up photos overnight” maps onto the real thing
| In the story | In the AI system |
|---|---|
| barcode scanner reading each box | system taking in structured input |
| comparison with the manifest | predefined rules or checks applied step by step |
| flagged missing or extra items | output such as alerts or exceptions |
| supervisor reviewing flagged cases | human oversight in an automated process |
A worked example
In a conference services office in Arusha, a conference support assistant receives an online room request for a meeting due to begin the next morning.
Starting from: The form specifies the meeting time, expected attendance, room layout and required equipment.
- The workflow checks whether every required field has been completed.
- It compares the requested time, capacity and equipment with the booking system's records.
- If all conditions are met, it creates a provisional booking and sends a confirmation. If information is missing or a conflict appears, it routes the request to the assistant.
- It records the outcome and schedules a reminder before the meeting.
Result: Complete requests that meet the programmed conditions are processed without repeated manual entry. Exceptions remain with the conference support assistant, rather than being guessed at or resolved by the automation.
In a United Nations workplace
In a liaison office, a training officer uploads a statement for delivery that must be cleared before close of business. The workflow sends the draft through the agreed clearance route, records each response, reminds reviewers with outstanding actions and alerts the front office when every required clearance has been recorded.
This is automation because the system executes predefined routing, reminder and status rules after the draft is submitted. It does not assess the statement's substance or treat silence as clearance unless that instruction has been explicitly configured.
Before you try thisThe draft must remain in a system approved for its sensitivity and must not be copied into an unapproved external tool. Before release, the training officer must confirm that the correct version reached the correct reviewers and that each required clearance was actually given.
Where you may hear it
- An ICT governance board reviewing a proposed clearance workflow
- A procurement unit testing approval routes during a vendor demonstration
- A country office workplanning session on recurring reporting tasks
- An internal audit meeting examining automated controls
Avoid confusion
Avoid confusion
Often confused with
Artificial intelligence
An umbrella term for computational systems using symbolic reasoning, search or learned models for tasks such as prediction, classification or content generation.
AI agent
An AI system that can pursue a goal by selecting and carrying out multiple steps, often using software tools, with oversight matched to its risks.
Machine learning
A way of building computer systems that learn patterns from examples so they can make predictions, classifications or recommendations on new cases.
Software
A set of coded instructions that tells a computer how to take input, process it and produce output.
Common misconceptions
Common misconception: Automation always means artificial intelligence.
Correction: Many automated systems do not use AI at all. A payroll run, an email filter based on fixed rules, or a door that opens when a sensor is triggered can all be automated without learning from data. AI is only one possible way to automate part of a task.
Common misconception: If a task is automated, no human is needed anymore.
Correction: People are still needed to design the process, check quality, handle exceptions, maintain the system and take responsibility for outcomes. In higher-risk settings, human review may be essential because an automated process can repeat the same mistake at scale.
Common misconception: Automation makes a process objective and error-free.
Correction: Automation can make a process consistent, but consistency is not the same as correctness or fairness. If the rules are badly designed, or the input data is wrong, the system can produce the wrong result every time and much faster than a person would.
Quick check
Which example is automation but not necessarily artificial intelligence?
A fixed spreadsheet formula is a classic example of automation: the steps are predefined and repeated automatically. The other examples usually involve machine-learning models, which are a subset of AI. Automation is the broader category.
Go deeper
Go deeper
The simple version
Automation means getting a machine or software to do repeated steps automatically after people set up the process.
The technical version
Automation is the execution of a defined workflow by machines or software with reduced direct human intervention. Inputs can include digital records, sensor signals, forms or messages. The system then applies programmed logic, which may be deterministic rules, scripts, thresholds, scheduling, orchestration across systems, or, in some cases, outputs from a machine-learning model. It then produces an action or result, such as updating a database, sending a notification, classifying an item, generating a report or triggering a handoff to a human. Automation can use AI but does not require it: rule-based systems, robotic process automation, industrial control systems and workflow engines can all automate tasks without learning from data.
Reviewed external resources
Primary sourceRecommendation on the Ethics of Artificial Intelligence (opens in a new tab)UNESCO
The UNESCO Recommendation helps readers who need a governance and ethics lens, especially policy and leadership audiences considering human oversight and responsibility.
Intermediatelink checked 2026-07-31
EncyclopaediaAutomation (opens in a new tab)Wikipedia
Wikipedia gives a broad overview of automation and its main uses, which suits readers who want a quick starting point.
Intermediatelink checked 2026-07-31
ArticleMachine Learning Glossary (opens in a new tab)Google
Google’s machine learning glossary helps readers distinguish automation from machine learning and related AI terms, which suits beginners building vocabulary.
Intermediatelink checked 2026-07-31
VideoVideo explanations: Automation (opens in a new tab)YouTube search
The YouTube search results page offers a range of video explanations on automation from different creators, which suits readers who prefer to compare several visual introductions.
search results, not one pagelink checked 2026-07-31
DiscussionCommunity discussion: Automation (opens in a new tab)Reddit search
The Reddit search results page shows community discussions and questions about automation, which suits readers who want to see how people use the term in practice but should be read critically.
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
Automation carries out a defined workflow with less direct human action. AI may be used in one step, but automation alone does not provide judgement, accuracy or accountability.