AI Agents in the Workplace: Who Gives Them Permission to Act?

As technology continues to advance, software providers like Salesforce and SAP are rapidly integrating artificial intelligence (AI) agents into their platforms. These AI agents can analyze data, make decisions, and even execute actions independently.

A recent example from a fast-growing enterprise shows the potential risks of this trend. When asked about the approval process for an AI decision to grant a $20,000 contract discount, the vice president was stumped, along with the director of IT and chief risk officer. They had spent months evaluating software licenses and security protocols but hadn't considered who gave the AI permission to make such decisions.

This oversight is a common issue in corporate governance, according to Gartner's latest adoption forecasts. By 2026, 80% of enterprise applications will deploy embedded AI capabilities. However, many companies are not adequately prepared for the risks associated with these autonomous agents.