artificial intelligence will force businesses to put their data in order / Economic news of Krasnoyarsk and the Krasnoyarsk Territory / Newslab.Ru

Today, 88% of companies have already tried to implement AI agents, but only about 10% were able to scale these solutions.

This gap is directly related to the lack of order in data, processes and infrastructure. Without their strict regulation, even successful pilots turn into an expensive experiment, said Sergei Golitsyn, head of the T1 AI department of the T1 IT holding, at the Data Fusion forum.

“AI is emerging as an indicator of digital maturity. EIf data is collected and stored haphazardly, processes are not described or executed, no intelligent agents will produce sustainable results,” emphasized expert.

According to him, companies will be forced to quickly unify IT systems.

“To create a large business based on artificial intelligence, you need a unified environment in which you can test hypotheses, train and validate ML models and then put them into production,” noted Sergey Golitsyn.

Expert reminded that similar platforms are already operating in the financial sector, telecoms and retail. NBut now the same level of maturity is required for generative models and intelligent agents, which must be built into a common platform rather than existing as a set of disconnected pilots.

“You don’t need to buy a new LLM, but something that will show you the result of its use – a platform that transparently increases the utilization of resources, increases the efficiency of models and the payback of projects,” – said He.

Sergei Golitsyn added thatAgainst this backdrop, a new trend is emerging: a more rigorous selection of tasks for AI systems.

“When infrastructure and computing are expensive, businesses have to decide much more carefully which tasks make sense to outsource to artificial intelligence. Without assessing the effect, even technologically mature projects become too expensive,” concluded Sergey Golitsyn.