The treasure hidden in the old emails: AI companies are rushing towards a new market

Companies may receive hundreds of thousands of dollars, and in some cases even millions, for their databases of emails, documents and work products. The buyers are artificial intelligence companies, and their goal is to teach the systems how professionals perform tasks, make decisions and solve problems over time.

The difference lies in the information that records the path to the result. A final legal document shows the wording chosen, but the sequence of correspondence and drafts that preceded it can show what questions were raised, what changed during the negotiations and how decisions were made. In an accounting firm, the work documentation may show how data was checked, when a problem was identified and what actions were required to resolve it.

This demand is accompanied by significant investment plans. Micro1 plans to invest a billion dollars in acquiring information from companies within a year. According to the forecast of the research company Gartner, the global expenditure on information for AI is expected to increase from about 826 million dollars in 2025 to about 6.5 billion dollars in 2027.

From an economic point of view, this development turns databases that until recently were mainly a storage cost into a potential asset. Alongside the AI ​​companies themselves, intermediaries are growing that locate companies that hold valuable information, examine its quality and manage transactions to sell it. Information that has been accumulated for years as part of the business’s current activity may suddenly have an independent value.

But the same documentation may also include details about employees and customers, business information and correspondence written for a completely different purpose. Selling it for training purposes raises questions of privacy, ownership and consent. Processes for removing identifying details can reduce some of the disclosure, but do not necessarily solve all the questions concerning the use of the information.

Alongside the race for chips and computing power, a market is also developing for the accumulated experience of businesses and employees. The emails, documents and work products may become another economic asset of the company, alongside the patents and software it has developed. Yaakov explains the demand for this knowledge as follows: “To reach the next level, the systems need to understand much better how the work itself is carried out.”