Scale AI’s Shift Aims to Solve ROI Challenges in Artificial Intelligence Investments

The challenge of achieving a return on investment in artificial intelligence remains: according to MIT, 95% of companies testing AI do not see the expected profits. In this world, the ambitions of tech leaders collide with reality – things won’t happen on their own, even if you connect a powerful AI model to business processes.

“There was a general promise: allegedly you could simply plug in an AI model… and everything would work,”

Scale AI acts as the engine for training artificial systems: large volumes of data are needed, which must be labeled and classified so models can distinguish a cat from a fish or differentiate different objects in images. Meta bought 49% of Scale’s shares in June for $14.3 billion, and the company’s overall valuation rose to about $29 billion. Founder Alexander Wang and several top executives left Scale to join Meta.

Although Scale continues to earn revenue from data labeling, new CEO Jason Droge shifted attention to a less visible but critically important area: helping other companies create their own data sets and develop tools to automate processes. He hopes to change the stereotype that AI deployments have no budget or profit.

“It will take years for large companies to implement AI tools that will be widely useful, generate revenue, and reduce costs,”

Despite the activity of industry leaders and the push to ride the AI wave, many companies still do not see the returns on their investments. MIT also notes that the U.S. economy continues to be pulled along by the growth of the AI market, but signs of a possible ‘bubble’ in the market are emerging due to uncertainty about stable profits from deployments.

Among Scale’s clients are Mayo Clinic, the Government of Qatar, Cisco, and Global Atlantic Financial Group. Last month Scale signed a contract for $99 million with the U.S. Department of Defense to develop AI applications for the Army. In the context of rising competition from Amazon and Microsoft, experts note that the most effective solutions are those that address concrete tasks and involve experts from the relevant industries, rather than universal AI tools.

Experts know that AI implementation can take years, but the one who can clearly define the problem and bring in external expertise can gain significant value. In such projects, having clear goals that enable increasing revenue, reducing costs, and optimizing business processes is the main challenge for many companies in the modern AI market.