Back to Glossary
BasicsAI Glossary

Algorithm

Quick Answer

An algorithm is a precise, step-by-step procedure for solving a problem or performing a computation. In the context of AI, algorithms define how a model learns from data, how it makes predictions, and how it optimises its performance.

In Depth

What Algorithm really means

Classical algorithms are fully specified by humans — for example, a rule that says 'if total order value exceeds £500, apply free shipping'. AI algorithms, by contrast, are typically learning algorithms: they define a procedure through which a model's parameters are updated from data.

Not all algorithms are AI. Sorting a spreadsheet, routing a delivery or compressing a file all rely on algorithms without any learning. Conflating the two often leads to misunderstandings about what AI can and cannot do.

Why It Matters

Business relevance for UK organisations

Understanding whether your organisation actually needs AI — or just a well-designed deterministic algorithm — can save significant time and money. Many supposed AI problems are better solved with simple business rules.

Real-world example

How this shows up in practice

A Glasgow accountancy firm replaced a proposed 'AI matching engine' with a deterministic rule-based algorithm, cutting development cost by 70% while achieving higher accuracy.