What is cargo cult (in tech)?

Cargo cult is the practice of reproducing the outward form of someone else's method, ritual or initiative without understanding the mechanism that makes it work, on the belief that copying the gestures is enough to copy the result. The name comes from the cargo cults that arose on some Pacific islands after World War II: local populations had watched, during the conflict, military aircraft arrive loaded with goods they had never seen before; once the war ended they built runways, control towers and mock radios out of bamboo, imitating what they had observed in the hope that the planes would return. Richard Feynman picked up the image in his 1974 Caltech commencement address "Cargo Cult Science", to describe research that follows the motions of the scientific method without the rigor that justifies them. Eric Raymond's jargon file later coined "cargo cult programming": code copied from another program and kept out of habit, with nobody left knowing what it is actually for.

Two enterprise shapes of the same mistake

In management the most common shape is surface-level agile: a team adopts the events of scrum, the daily, the board, two-week sprints, without the constraint that makes them useful, namely empiricism, the real willingness to change course based on what has been learned. The ritual stays, decisions still arrive from outside, and after a few cycles the team stops even declaring a sprint goal because nobody takes it seriously anymore. The second shape, more recent, concerns AI adoption: a company replicates a competitor's initiative, the same model, the same vendor, the same press release, without the clean data, governance and internal skills that made the original project work. It is distinct from agent washing: there a vendor labels something an agent that is not one, here it is the buyer copying someone else's outcome without the conditions that produced it.

The small manufacturer that copies the competitor's chatbot

A manufacturing company of about forty people reads that a competitor launched an AI assistant for customer service and decides to do the same within the quarter. It buys the same platform, copies the prompt found in a webinar, and connects it to the product catalog as it stands, an Excel sheet updated at irregular intervals by three different people. The competitor had invested a year earlier in cleaning up the catalog and in a person who kept it consistent: the copied initiative worked because of that, not because of the chatbot itself. The new assistant answers with expired prices and discontinued sizes, customers call the front desk anyway, and the company concludes that "AI does not work for our industry", when the problem was the underlying data, never touched.

Why it matters for decision makers

Cargo cult is costly because it moves the budget toward the visible part of an initiative, the ritual, while the part that actually generates value is almost always invisible from the outside: data discipline, governance, the people who know why a rule exists. Before replicating someone else's method or initiative it is worth asking what, exactly, makes it work where it originated, and whether that condition exists here too; if the answer is no, copying the form produces only the cost, not the result. It is the same filter that applies to an AI project at risk of failing for similar reasons, listed in AI project failure causes.