COMPEL Glossary / algorithm
Algorithm
An algorithm is a set of step-by-step instructions or mathematical rules that a computer follows to solve a problem or complete a task.
What this means in practice
In AI, algorithms are the procedures that enable models to learn patterns from data -- for example, gradient descent is the algorithm that adjusts neural network weights during training. For non-technical transformation leaders, the key insight is that different algorithms are suited to different problems: a decision tree algorithm works well for simple classification, while a transformer algorithm powers large language models. Understanding algorithms at this level helps leaders ask better questions during vendor evaluations and technology selection. Algorithmic accountability -- the question of who is responsible when an algorithm makes a harmful decision -- is a central governance concern in the COMPEL framework.
Why it matters
Understanding that different algorithms are suited to different problems helps leaders ask better questions during vendor evaluations, technology selection, and AI strategy decisions. Non-technical leaders who grasp algorithmic basics can distinguish between genuine capability and vendor hype, make more informed investment decisions, and participate meaningfully in governance discussions about algorithmic accountability and transparency.
How COMPEL uses it
Algorithmic understanding is part of the AI Literacy domain (D3) in the People pillar, assessed during Calibrate to determine baseline organizational knowledge. The Organize stage designs literacy programs that build appropriate algorithmic understanding at each organizational level. Algorithmic accountability — the question of who is responsible when an algorithm causes harm — is a central governance concern addressed in the Governance pillar during Model and enforced during Evaluate.
Related Terms
Other glossary terms mentioned in this entry's definition and context.