52% of organizations already use AI agents in daily operations, Google Cloud Israel executive says

In an interview during Google and Calcalist’s AI Week, Calcalist’s Haggai Gilboa spoke with Alit Ben-Bassat Noriel, marketing vice president at Google Cloud Israel, about the shift from task-based AI to AI agents and how work will change in the coming years.

Ben-Bassat said the industry is at “the most dramatic turning point since generative AI entered our lives.” Citing a global survey of 3,500 senior executives, she said 52% of organizations already using generative AI have also embedded AI agents into day-to-day operations. She described three stages of maturity: simple chatbots that retrieve information and perform basic actions, dedicated agents for functions such as service or creative work, and the most advanced stage, multi-agent workflows, where several AI systems communicate and run end-to-end processes.

According to her, organizations that first define a clear business need for an agent are already seeing positive ROI. She pointed to applications in customer service and experience, sales and marketing, cybersecurity, and technical support. As an example, she said salespeople who once prepared for a client meeting for three hours can now use a smart agent connected to the organization’s APIs to build a presentation in five minutes. In cyber, she added, since attackers use AI, defensive teams must operate at “machine speed.”

She offered three recommendations for companies starting out: leadership must define the business need, understand the technology, set expected ROI, and allocate dedicated budgets; organizations should begin with repetitive, time-consuming processes rather than trying to build a full architecture immediately; and they must establish strong data governance and security, with secure, controlled access to core systems and a zero-trust framework.

Looking three years ahead, Ben-Bassat predicted a hybrid workplace where people manage workers alongside agents, with every employee having a personal AI agent or a network of agents. She said the emerging skill that will matter most is orchestration, meaning precise goal-setting, context management, guardrails, and critical thinking. Human workers, she said, will become less hands-on and more strategic, because “the human capital is not becoming less important, it is becoming more strategic.”