Businesses are rushing to use AI agents, but safety is lagging behind
A new Deloitte report finds that while most companies want to customize AI agents for their needs, only a small number have set up strong safety checks.
Out of 3,235 leaders surveyed across 24 countries, 85% plan to tailor these tools, but just 21% have solid oversight in place.
AI agents are spreading fast—with real risks
Right now, about a quarter of companies use AI agents regularly, and that's expected to jump to nearly three-quarters in two years.
These bots aren't just answering questions—they're signing documents and making purchases on their own.
But with this freedom comes the risk of unpredictable actions and security threats like prompt injection attacks.
The price tag for all this AI? It's growing fast
AI costs are rising in tech budgets—some firms report massive monthly bills.
To keep things under control, leading companies are mixing cloud and local systems and making sure their AI projects actually deliver results they can measure.