$547 billion in AI dollars gone. Recommendation: take a communication course.

An employee at a data company, a company that sells data infrastructure, has written an opinion piece about why AI fails. The answer is culture. Not the technology, not the data, not the parties that have been pocketing billions over the past three years for systems that do nothing. Culture. You’re just not talking to each other properly.

Eighty percent of all bridges collapsed. The crane supplier showed up in a trade publication. The problem: architects and contractors don’t speak the same language. Buy our cranes and we’ll train your teams. He’s speaking in San Francisco next month.

You know how this works. Data quality is the primary failure factor. Data quality infrastructure is the product. The employees who shout that something isn’t working are the first to be passed over. Then you write an article about cross-functional playbooks, send it off, and the invitation takes care of itself.

A direct attack on technology is contestable. A soft message about collaboration isn’t, because who’s against collaboration? The people who lost their jobs when the pilot was scrapped? They don’t give keynotes. They don’t write opinion pieces. They don’t appear in the footnotes of the studies proving the system fails. They’re the evidence, which is exactly why they’re absent from the analysis.

The system produces failure rates. Then a market for explaining those failure rates. Both markets served by the same parties, invoiced to the same organizations, who’ll kick off another pilot next quarter.