Control is lagging behind: only half of companies have tools to monitor AI
25 February 20:20
The new Microsoft Cyber Pulse report notes the rapid spread of AI agents in large businesses: four out of five large companies already use them in their daily work, according to "Komersant Ukrainian".
At the same time, management and control systems often cannot keep up with the speed of scaling.
AI agents are going beyond IT departments
According to the company, more than 80% of Fortune 500 companies are deploying active agents created using low-code and no-code platforms, including Fabric, Foundry, Copilot Studio, and Agent Builder.
Development is no longer limited to technical specialists. Agents are created and implemented by employees from various departments, from sales to finance. They are used for:
- preparing commercial proposals,
- analyzing financial data,
- processing security alerts,
- automating routine processes.
The largest number of active agents are concentrated in the EMEA region (42%). Next are the US (29%), Asia (19%), and other countries in the Americas (10%). By industry, the leaders are technology and software development (16%), manufacturing (13%), financial services (11%), and retail (9%).
The risk of “shadow AI”
The report warns that the rapid deployment of agents comes with risks. They often gain excessive access to data, and their use is not always transparent to management.
This creates the phenomenon of so-called “shadow AI” — tools that employees use without consulting IT services. According to Microsoft estimates, 29% of employees already use such agents in their work.
Renate Strazdina, National Technical Director of the Northern Europe cluster at Microsoft, noted:
“AI agents are becoming digital colleagues, often outpacing the ability of organizations to implement proper governance.”
She emphasized that for Ukraine, one of the main targets of cyberattacks, innovation must develop alongside Zero Trust principles, centralized management, and constant monitoring.
Is business ready for control?
Only 47% of companies reported having special control tools for generative AI. Microsoft emphasizes that artificial intelligence management cannot remain solely the responsibility of IT or CISO.
Lawyers, compliance specialists, HR, and top management must be involved in the process.
These trends are particularly important for Ukrainian businesses. Companies in IT, finance, manufacturing, and the public sector are increasingly integrating AI tools. At the same time, constant cyberattacks are raising the bar for access control, data protection, and personal responsibility of management.
The spread of AI agents is already changing the way large companies operate. The next step is to determine who will control these “digital colleagues” and how.
