Global spending on artificial intelligence has reached $2.5 trillion, yet roughly 80% of corporate projects still fail to produce the promised business outcome. That was the central message in an interview on RBC Radio with Alexey Shpilman, Head of AI Technology Development at T-Technologies.
Shpilman said companies should separate AI initiatives into three categories: employee tools, process-optimization solutions, and new higher-risk directions that may bring the biggest payoff. In his view, each type requires a different approach to implementation and measurement.
He added that internal AI services are already widely used inside the company, with about 90% of employees relying on them. In T-Bank, AI now handles 45% of customer support requests, but the goal is not headcount reduction; instead, the focus is on service quality and customer satisfaction.
Shpilman also said AI is faster than humans and can know more, but it still makes mistakes and therefore needs human oversight. He pointed to the need to identify areas where Russia can gain an advantage in the AI race and to preserve human interaction where it remains essential.